Nvidia once again posted a $2 billion beat to revenue estimates, reporting YoY growth of nearly 94% to over $35 billion in revenue. Data center revenue more than doubled in the quarter to over $30 billion, speaking volumes as to the level of demand for its GPUs given that Blackwell had not begun to ship in Q3.

Blackwell matters – a lot. As stated in the webinar, the more the team looked at the details of the release, the clearer it has become that 2025 will be Nvidia’s year – again. Some of this was covered in the pre-earnings write-up published this morning.

On the positive side:

• Hopper drove the beat that analysts were expecting, with UBS tagging the beat at $2 billion. This did, indeed, materialize in the earnings report. This beat is clearly not lackluster but given the performance of the stock, up about 850% since two years ago when Hopper began to ship, the product cycle is lackluster and not able to reinvigorate the stock.
• My primary message going into tonight’s results was that the I/O Fund is tracking supply chain signals indicating the new generation of GPUs shipping in full volume by mid-2025 (and beginning to ship in the January quarter) will far exceed the GPU sales we saw in 2023 and 2024 combined. This was echoed tonight when Jensen Huang stated: “You see now that at the tail-end of the last generation of foundation models were at about 100,000 Hoppers. The next generation starts at 100,000 Blackwells.”
• The importance of big tech capex was also echoed with the CEO stating we will see $1 trillion in data infrastructure rebuild before he expects to see digestion from the hyperscalers. As Damien on the team helped to point out last week, we are at a quarter-trillion right now. Per the CEO: “I believe that there will be no digestion until we modernize a trillion dollars with the data centers.” That would imply another 3X from here for the remaining three-quarter trillion – not in stock price, but in capex. Presumably, it would mean a higher trajectory for the stock price in terms of valuing that revenue.
• Management debunked the supply chain rumors (which the inaccuracy is getting to be a tad annoying at this point). When asked about supply chain rumors, the CEO stated “Blackwell production is in full steam. In fact, as Colette mentioned earlier, we will deliver this quarter more Blackwells than we had previously estimated.” As expected, Colette Kress was tight lipped and no official number was provided. I have personally found Nvidia’s management style to be to our benefit as they were a closed book two years ago in the quarters that preceded the historic ramp.

What the Street Asked About:

• Despite the rather large top-line beat, margins were relatively in line with guidance, and forecast to contract nearly 2 points sequentially.
• Never underestimate Wall Street’s ability to miss the bigger picture. Analysts on the call cross-examined this 200 bps decline despite Nvidia having an operating margin of over 60% compared to most of the Mag 7 having operating margins at half that. The CFO was clear that following Blackwell, the gross margin will eventually return to its current percentage: “As Blackwell ramps, we expect gross margins to moderate to the low-70s. When fully ramped, we expect Blackwell margins to be in the mid-70s. GAAP and non-GAAP operating expenses are expected to be approximately $4.8 billion and $3.4 billion, respectively.”
• Supply chain constraints: There has been some FUD published by The Information back in August and again this week. Management provided a strong comment to refute these claims, primarily that: “We completed a successful mask change for Blackwell, our next Data Center architecture, that improved production yields.” Yields is what matters here and this comment along with Q4 seeing more Blackwell revenue than previously estimated helps to eliminate these concerns.
• Broadly speaking, there are supply constraints but this is nothing new as it’s been widely understood Blackwell is already sold out for next year.  
• As we close out the year and move into 2025, investors should be prepared to hear about China and tariffs. Per the pre-earnings report, Nvidia has limited exposure at 12.5% yet it’s quite clear with weak SMH and SOXX ETF price action that the market is pricing in this impact. It’s unclear to me today how TSM will be viewed in terms of tariffs given the Arizona plant is up and running. You can view our webinar clip here regarding SMH.

Fiscal Q3 2025 Results:

As stated, Hopper drove the beat that analysts were expecting, with UBS tagging the beat at $2 billion. However, due to declining from peak revenue growth of 265% earlier this year, Hopper-driven growth of 94% is not what will drive the stock up for the next leg higher. Nvidia investors, such as myself, will need Blackwell’s pricing power and Blackwell’s clear demand signals to re-invigorate the stock.

As stated, the one weak link of the report was Q4’s margin guidance, with management pointing to potential contractions down the line as Blackwell ramps.

Revenue

Nvidia reported 93.6% YoY growth to $35.08 billion in revenue, well ahead of the consensus estimate for $33.13 billion (83% YoY). Nvidia is now lapping its peak growth quarters, Q3 FY24 to Q1 FY25, where revenue more than tripled each quarter as Hopper ramped tremendously fast. Management said in Q3 that the H200 “grew significantly in the quarter.”

Growth technically is decelerating nearly 30 points in Q3 and growth will further decelerate nearly 24 points next quarter, but to be reporting above 93% YoY and almost 70% YoY versus 200-260%+ growth comps is a strong report to say the least.

For Q4, management guided for revenue of $37.5 billion, +/- 2%, just slightly ahead of consensus estimates for $37.02 billion at the midpoint. Management noted that they have “completed a successful mask change for Blackwell…that improved production yields. Blackwell production shipments are scheduled to begin in the fourth quarter of fiscal 2025 and will continue to ramp into fiscal 2026.”

Both Hopper and Blackwell will be shipping in tandem, placing more emphasis on supply constraints moving forward, as management was clear in saying that both products have “certain supply constraints” with Blackwell’s demand “expected to exceed supply for several quarters in fiscal 2026.”

China revenue was 15.4% of revenue compared to 12.7% year-to-date. This is down from the low-20% range last year.

Key Segments

It should be of no surprise that data center revenue beat estimates in the quarter, but what’s interesting is that the segment posted the largest surprise relative to estimates since Hopper’s breakout quarter in FY24.

Data center revenue of $30.77 billion increased 112.0% YoY and 17.1% QoQ, beating estimates by $1.95 billion. Assuming a similar mix as the current quarter, Q4’s data center revenue would be implied to be nearly $32.5 billion.

In the segment, data center compute revenue was $27.64 billion, rising 132% YoY and 22% QoQ. Networking revenue increased 20% YoY but declined (15%) QoQ to $3.13 billion – this slowed sharply from 114 % YoY growth in Q2.

Management said networking growth was driven by Ethernet for AI; “NVIDIA Spectrum-X Ethernet for AI revenue increased over 3 times year-on-year and our pipeline continues to build with multiple CSPs and consumer Internet companies planning large cluster deployments.” It was also indicated that networking would resume sequential growth next quarter: “So this quarter is just a slight dip down and we're going to be right back up in terms of growing. They're getting ready for Blackwell and more and more systems that will be using not only our existing networking but also the networking that is going to be incorporated in a lot of these large systems that we are providing them to.”

• Gaming revenue of $3.28 billion increased 15% YoY and 14% QoQ, driven by GeForce RTX series 40 GPUs and game console SoCs.
• Pro Viz revenue of $486 million increased 17% YoY and 7% QoQ, driven by the ramp up of RTX GPU workstations.
• Automotive revenue of $449 million increased 72% YoY and 30% QoQ, accelerating 35 bp QoQ from 37% YoY growth in Q2, driven by Nvidia’s self-driving platform.
• OEM and other revenue of $97 million increased 33% YoY and 10% QoQ.

Margins

Despite the rather large top-line beat, margins were relatively in line with guidance, and forecast to contract nearly 2 points sequentially. This forecasted weakness as Blackwell ramps may be one of the factors behind the initial post-earnings sell-off, with GAAP operating margin seen coming back towards 60%.

• GAAP gross margin was 74.6%, just ahead of guidance for 74.4%. Adjusted gross margin was 75%, in line with guidance. This reiterated our view from last quarter that Q1 was the peak for gross margins, as margins have contracted about 380 bp since then.
• For Q4, management guided for GAAP gross margin of 73%, +/- 0.5%, and adjusted gross margin of 73.5%, +/- 0.5%, for a sequential contraction of ~150-160 bp.
• GAAP operating margin was 62.3% in Q3, increasing slightly from 62.1% in the prior quarter but up from 53.1% in the year ago quarter. Adjusted operating margin of 66.3% dipped slightly from 66.4% in Q2, but increased from 64.8% in the year ago quarter.
• For Q4, similar to gross margins, management guided for sequential contraction based on operating expense forecasts. GAAP operating margin is implied to be 60.2%, while adjusted operating margin is implied to be 64.4%, or about a 200 bp sequential contraction. 

• GAAP net margin was 55.0%, down from 55.3% last quarter but up from 51.0% in the year ago quarter. Adjusted net margin was 57.0%, up from 56.4% last quarter and 55.3% in the year ago quarter.

  While it may seem like a small difference, putting it to the scale of revenue growth will show that net income has more than doubled YoY – GAAP net income was $19.31 billion in Q3, up from $9.24 billion last year despite only a 4-point margin expansion.

• GAAP EPS of $0.78 beat estimates by $0.08, and represented YoY growth of 111%. Adjusted EPS of $0.81 beat estimates by $0.06 and represented YoY growth of 103%.

Cash and Balance Sheet

Cash flows remained strong in the quarter, with operating and free cash flow margins both expanding sequentially.

• Operating cash flow was $17.63 billion, rising 141% YoY and 22% QoQ. OCF margin was  50.3%, expanding from 48.2% last quarter and 40.5% last year; to note, this remains below the 58.9% margin from Q1.
• Free cash flow was $16.79 billion, rising 138% YoY and 25% QoQ. FCF margin was 47.9%, up from 44.9% last quarter and 38.9% in the year ago quarter.
• Inventories totaled $7.65 billion, increasing nearly 60% YoY and more than 14% QoQ. Purchase commitments and obligations for inventory and capacity also rose 4% QoQ to $28.9 billion. Capacity and supply pre-payments were $5.2 billion, reaffirming that Nvidia is well prepared to launch Blackwell in full-force.
• Cash and equivalents totaled $38.49 billion, while debt totaled $8.46 billion.

Earnings Call:

To elaborate on the margin concerns, here was an exchange in the Q&A:

Timothy Arcuri:

“[…] And then Colette, you kind of talked about Blackwell bringing down gross margin to the low-70s as it ramps. So I guess if April is the crossover, is that the worst of the pressure on gross margin? So you're going to be kind of in the low-70s as soon as April. I'm just wondering if you can sort of shape that for us. Thanks.”

Colette Kress

Sure. Let me first start with your question, Tim. Thank you regarding our gross margins, and we discussed our gross margins as we are ramping Blackwell in the very beginning and the many different configurations, the many different chips that we are bringing to market, we are going to focus on making sure we have the best experience for our customers as they stand that up. We will start growing into our gross margins, but we do believe those will be in the low 70s in that first part of the ramp. So you're correct, as you look at the quarters following after that, we will start increasing our gross margins and we hope to get to the mid-70s quite quickly as part of that ramp.”

–End Quote

For More Reading:

Please reference our pre-earnings write-up which summarizes my current thoughts on the stock.

Conclusion:

I said on Fox Business News on Tuesday that I would love to get Nvidia lower, and I truly would. The nitpicking around the margins, the weaker semiconductor peers, the low volume as the stock trades near its all-time highs, the tariff concerns … one or all of these may present us that opportunity.

Nvidia’s fundamentals are a perfect 10. The pre-earnings report had stated: “Make no mistake, Nvidia is the best stock of the decade and we are only four years in. The big picture is that Nvidia's trajectory will continue due to two words: pricing power.”

We are already tracking a 30% minimum difference between GB200NVL72 orders and what Wall Street has estimated for next year. When you add that the DGX B200 systems will be priced 40% higher, and if we assume pricing power affects more SKUs the way it’s going to affect the DGX B200 systems, then we could see about 70% upside next year for Nvidia. Now, the I/O Fund likes to be aggressive, it’s why you’re here. If we can get the stock lower, that potential upside increases.

Wish us luck – and keep an eye on those trade alerts!

Recommended Reading:

• Astera Labs: Hypergrowth AI Networking Stock
• Chainlink: A Front Runner Among Blockchain Technologies
• Real Vision Video Interview: Will Nvidia Continue to Dominate AI?
• AMD Q3 2024: GPU Revenue at 22%, and AI PCs have a Leader

Astera Labs is a stock that simply cannot be ignored despite having a fraction of the market cap compared to much larger AI hardware players.

The company offers a few key products that are enabling larger and faster AI clusters. For data center AI accelerators, the company was first to offer PCIe 5 switches and retimers. Growth will likely continue due to next-generation PCIe 6 products with higher average sales prices that will be released in 2025 and ramp in 2026.

There are additional growth opportunities, including a software architecture that allows hyperscalers to monitor their data center infrastructure and increase utilization rates; the software helps to incentivize hyperscalers to use Astera Labs hardware products. The company also offers ethernet smart cable modules and is introducing a new CXL product for CXL-enabled CPUs next year.

We are interested in Astera Labs for the increased average sales prices that are expected to persist at least through 2025-2026 due to the Aries products and upcoming Scorpio products. Astera Labs is a highly technical company, yet this quote clearly communicates why the growth trajectory can sustain:

“One is generally speaking with each new generation of a protocol like PCIe going from Gen 5 to Gen 6. There is an ASP uplift. That's number one. Number two, of course, we were hinting at the Scorpio product line, which because of the value it delivers to customers is at a higher ASP, as you can imagine. So overall, if you look at the design wins we have today, the dollar content per GPU goes up, that's one way to look at it based on what we've shared before […] So overall, if you look at sort of the increasing speed, additional product line as well as the fact that the internally developed platforms, AI accelerated database platforms, they are starting to gain more and more traction. So when you look at all of them, on an average, our content is on the up.”

Investment Thesis:

• Current Growth Driver: Increased average sales prices are being driven by CPUs, GPUs and ASICs all moving to the new PCIe 5.0 standards. Arista Labs’ Aries Retimers and PCIe 5.0 components is driving the current growth, and the company is unchallenged in this new generation of PCIe, which came to market for AI accelerators only recently with Nvidia’s H200s.
• Catalyst: Scorpio is a new product that is expected to expand the TAM to more than $12B by 2028: Astera Labs is releasing a PCIe Gen 6 fabric switch custom designed for AI data flows with high performance per watt compared to incumbents. Rather than building a large switch, the company built a smaller device that is more efficient for high-speed signals. With Scorpio, Astera Labs is defending its dominance in PCIe5 by doubling the bandwidth at lower power requirements than the 5th generation of PCIe.
• The market demand for Astera Labs is healthy, as a major supplier to Nvidia’s PCIe-enabled GPUs (note: most of Nvidia’s GPUs use their in-house NVLink). Yet, Arista also supplies other AI accelerator platforms for Big Tech, and Arista is the only provider for PCIe5. In addition to hardware, ALAB offers Big Tech data centers software-defined architecture called COSMOS that allows for performance monitoring. There is some favorable vendor lock-in dynamics with COSMOS as the many different product lines can be optimized and monitored with the software.
• Financials on Astera Labs are impressive as the company is hypergrowth at 206% growth last quarter and 153% guided next quarter. The free cash flow margin is at 41%. The company is not GAAP profitable due to being a recent IPO with outsized stock-based compensation. However, there is a path to GAAP profitability, which is detailed in the financials section below.

Overview of Astera’s Products:

Broadly speaking, Astera’s products are seeing increased relevance as AI clusters grow to support hundred-billion and trillion parameter models. The company’s hardware and software increase AI server performance and productivity for the current generations and future generations of AI accelerators.

Transition to PCIe5 and PCIe5 Retimers are Driving Astera’s Growth:

The Aries products offer PCIe5 interfaces that GPUs and AI accelerators (like custom silicon) use to connect components including ethernet networking. Compared to the previous generation, PCIe5 is twice as fast with data transfer rates that reach up to 128 GBs/s on multi-directional bandwidth in each lane. By increasing the data transfer for each lane, it allows more lanes to become available to help leverage the power of the GPU or AI accelerator.

PCIe5 Retimers are chips that boost signals across high-speed components and are seeing increased demand, starting with Nvidia’s H200s, and also for application-specific chips. Specifically, Aries Retimers and smart cable modules allow hyperscalers to connect multiple racks together with up to 7 meters of copper cables. Aries can also go up to 50 meters with optical fiber. According to a presentation at Nvidia’s GTC event, this is 3X the standard reach defined in PCIe specs.

Despite PCIe5 being out since 2019, it was Nvidia’s H200s released in 2024 that were the first data center GPUs to use PCIe5. What’s interesting is that Astera is said to have captured 95% of the XPU market, which refers to application-specific chips that are specific on a product level. Per an analyst on the earnings call: “Majority of the XPU shipments are still going to be I think Gen 5 based where your market share is still somewhere in the range of, I think, like 95%.”

Management also stated: “The upside to the guidance was driven largely from Aries' revenue, both for the third-party GPUs, but also as well with the strong ramps on new platforms on the internally developed AI accelerators. And we're seeing that across multiple hyperscaler customers, so it's not just one. So the upside was largely driven by that Aries revenue.”

Notably, Aries devices are used to interconnect AI accelerators with CPUs and networking, yet are also used for backend networking between GPUs for larger clusters. Astera supplies the HGX H100 systems with PCIe-based GPUs with up to 25 retimer chips per HGX system.

Scorpio PCIe6 Custom-Built for AI is a Catalyst:

Scorpio is where the excitement is building for continued growth next year with management stating it will “exceed 10% of revenues in 2025” with “good momentum going into 2026.” This is due to the PCI Express switches being custom built for AI purposes, whereas in the past, PCIe was built for storage and then retrofitted for AI purposes.

PCIe 6 doubles the bandwidth from the 5th generation, with up to 256 GB/s of bandwidth per lane, which will require faster supporting components, such as the retimers that Astera Labs offers. The demo from GTC showed the Scorpio fabric switches (name released in October) delivering twice the bandwidth with less power at 11W instead of the 13W from the PCIe 5 interfaces.

There are two Scorpio fabric switches. The Scorpio P-Series is a small chip that connects the CPU, GPU, NIC and NVMe storage. Rather than building a large switch, the company built a smaller device that is more efficient for high-speed signals to help feed GPUs with data. The fewer ports and smaller switch decrease complexity in a bid to compete against Broadcom with twice the lane count.

The X-Series is for back-end networking in GPU-to-GPU configurations, and will offer a higher port count. Astera is essentially building something similar to Nvidia’s NVSwitch with the X-Series, but for PCIe-enabled GPUs: “And this one, like Mike noted, it's a greenfield use case, meaning if you keep Nvidia and NV Switch aside, everyone else is starting to build configurations that are obviously going to need some kind of a switching functionality, which is what we are addressing with our X Series device.” The X-Series improves efficiency for ever-increasing AI cluster sizes. The majority of AI clusters are in the tens of thousands GPUs, but are expected to go to the hundreds of thousands (already has with X and some other Big Tech companies), and will see AI clusters with millions of GPUs over the next couple of years.

Here is another quote from the management team as to why the X-Series Scorpio switch fabric is a big opportunity for their company:

“X series will have a bigger TAM. The TAM today is nearly zero. It's not very commonly used outside of the Nvidia ecosystem. We do expect many hyperscalers to start deploying this, starting with the X family and the designs for which that we have. And we are able to do that because of the architecture that we have. Because of our software-defined architecture, we can customize many parts of the X-Series to cater to the specific requirements of the hyperscalers both on the side of performance, the exact configuration that they require in count and so on and also the diagnostics framework that they require to monitor their infrastructure. So over time, we do expect X-Series to become larger [than the P-Series].”

COSMOS Software:

In order of importance, COSMOS Software ranks higher than some of Astera’s hardware as it offers performance monitoring for data center infrastructure. This is especially important for Big Tech companies concerned with utilization rates for expensive GPU systems.

The adoption of the software stack to monitor for performance is expected to increase with the Scorpio X-Series:

“Where the Scorpio family sits, we have access to a lot more diagnostic information. And we can couple that with the information that we are collecting from our other families deployed such as Aries and even Taurus to provide a holistic view of the AI infrastructure to the data center operators. So both from the hardware side, the kind of the purpose-built nature of these devices as well as the software stack that comes with it is a big differentiator for us.”

Taurus Ethernet Smart Cable Modules:

Astera Lab offers Ethernet smart cable modules which help to alleviate bandwidth issues with 100-gig per lane connectivity over copper cables including AEC. The company recently released 400-gig Ethernet SCMs, which help to stabilize the network. When thinking of ALAB’s investment thesis, Ethernet is not top-of-mind given the sheer size and lead we see from Broadcom, Nvidia’s Spectrum, and Arista Networks.

Right now, the maximum bandwidth supported by PCIe 5.0 is 400Gbps per port. By using 106Gbps PAM4 SerDes, ASICs can be tuned to support 100, 200 and 400 Gbps port speeds. To work around this, and to achieve 800Gbps, larger chip makers are building NICs directly into the accelerator. According to The Register, the 800Gbps ports built into accelerators may reduce bottlenecks before PCIe 6.0 arrives on the market. The larger Ethernet players are moving quickly on this, and we will need to keep an eye on Astera Labs to determine if the company’s Taurus product can remain competitive.

Leo Compute Express Link (CXL):

Leo is slated to impact revenue in 2025 when data center platforms plan to utilize CXL technology for memory bandwidth and capacity bottlenecks. Next year, CXL-capable CPUs will become broadly available. We’ve covered in the past how CXL is a new server architecture that “dynamically assigns memory resources between servers.” The result is boosted memory bandwidth and also at a lower cost than adding more CPUs. Partially-disaggregated racks are expected to deploy in 2024-2025 with separate compute, memory and I/O racks with the interconnect being CXL.

Per Astera’s management team: “In the past, this was done by adding additional CPUs into the server box to provide for more memory channels. But what we have demonstrated is that by using Leo you're not only able to get the higher performance by the added memory. But from an overall TCO standpoint, it's significantly less than adding additional CPUs. “

Financials: Strong Revenue Growth, Expanding Margins

Astera Labs reported solid top-line and bottom-line beats in the recent Q3 results. The company reported record revenue of $113.1 million, up 47% QoQ and 206% YoY, beating estimates by 16.1%. The adjusted operating margin expanded to 32.4% from 2% in the same period last year, which was helped by strong operating leverage. The adjusted EPS of $0.23 beat estimates by 35%.

Revenue:

The company is one of the fastest-growing tech companies and is emerging as a rising star in the AI data center networking space. The company’s Q3 revenue grew by 206.2% YoY and 47% QoQ to $113.1 million.

Jitendra Mohan, CEO and co-founder of the company, said in the Q3 2024 earnings call, “Our business has entered a new growth phase with multiple product families ramping across AI platforms, featuring both third-party GPUs and internally developed AI accelerators, which drove the Q3 sales upside verus our guidance.”

• Management also provided a strong Q4 revenue guide of $126 million to $130 million, representing YoY growth of 153.4% at the mid-point. This also represents QoQ growth of 13% at the midpoint.

  According to management, the QoQ growth is being driven by “our Aries product family across a diverse set of AI platforms, some of which are just starting to ramp and also from our Taurus SCM for 400-gig applications, and additional preproduction shipments of our Scorpio P-Series switches.”

• Analysts estimate revenue to grow 103.5% to $132.78 million in Q1 2025 and 85.2% YoY to $142.34 million in Q2 2025. While the growth rates are strong, growth is expected to slow down due to tougher comps.

Jitendra Mohan said, “Looking into Q4, we expect our revenue momentum to continue, largely driven by the Aries PCIe and Taurus Ethernet product lines. The Scorpio Fabric Switches are continuing to ship in preproduction volumes.”

The newly introduced Scorpio Fabric Switches are expected to increase the total market opportunity to more than $12 billion by 2028 for the company. Scorpio Switches are also expected to constitute more than 10% of revenue in 2025.

Source: Company website

Analysts expect 2024 revenue growth of 230.9% YoY to $383.14 million, followed by 55.5% and 40.4% in the subsequent years. Meanwhile, management comments seem to imply the growth recently reported will sustain, implying the growth phase has only begun: “Our business has entered a new growth phase with multiple product families ramping across AI platforms, featuring both third-party GPUs and internally developed AI accelerators, which drove the Q3 sales upside versus our guidance.”

Management also later stated: “Yes, right now, our visibility is very strong, both as always with our backlog position, but also the breadth of designs we have — right now, we're really kind of entering a new phase of growth here where our revenue streams are clearly diversifying […].”

Margins:

The company’s margins are improving, helped by strong operating leverage. However, the product mix might weigh on the margins going forward. Management mentioned in the recent earnings call that they have a long-term gross margin target of 70%.

• Q3 gross profits grew by 212.7% YoY to $87.88 million or 77.7% of revenue compared to 76.1% in the same period last year.
• This compares to 73.5% for FY2022 and 68.9% for FY2023.
• Adjusted gross margin was 77.8% compared to 76.1% in the same period last year.
• Management gross margin and adjusted gross margin guide for the next quarter are 75%. This is down sequentially due to higher product mix towards hardware solutions during the quarter.
• Operating margin was (-7.9%) compared to (-5.3%) in the same period last year. The adjusted operating margin expanded to 32.4% from 2% in the same period last year, helped by strong operating leverage.
• Management operating margin guide for the next quarter is (-4.3%) and adjusted operating margin guide is 32.4%.
• It’s important to consider that stock-based compensation is quite high due to the recent IPO at 40.3% this quarter, and was at 56% last quarter. Once SBC naturally levels out, this company has strong enough margins to become GAAP profitable.
• Net loss was (-$7.6 million) or (-6.7%) of revenue compared to (-8.5%) in the same period last year.
• Adjusted net income improved significantly to $40.28 million or 35.6% of revenue compared to (-$0.41 million) or (-1.1%) of revenue in the same period last year.

The difference between the GAAP and non-GAAP net income is due to high stock-based compensation. Stock-based compensation was $45.5 million or 40.3% of revenue in the recent quarter. Stock-based compensation has been lumpy as the company’s IPO was in March 2024.

EPS:

The company beat EPS estimates helped by solid operating leverage. Q3 GAAP loss per share was ($0.05) compared to ($0.08) in the same period last year, beating estimates by 28.1%. Adjusted EPS was $0.23 and beat estimates by 35%.

• Management has guided the Q4 GAAP EPS in the range of $0.04 to $0.06.
• Notably, the company will not be GAAP operating income positive next quarter as the guide suggests an operating loss of (-$5.5 million) and interest income of about $10 million in the next quarter will make it GAAP profitable. However, it’s very close to being GAAP profitable across all margins, and we think it’s only a matter of time before this happens.
• Adjusted EPS guide is $0.25 to $0.26.
• Analysts expect adjusted EPS to be $0.25 in Q1 2025 and $0.27 in Q2 2025.
• Analysts expect strong EPS growth. They expect 2025 adjusted EPS to grow 58.9% YoY to $1.14 and 48.6% YoY to $1.70 in 2026.

Cash Flow and Balance Sheet:

The company’s cash flows margins are high and are also improving due to increased profits.

• Q3 operating cash flow was $63.5 million or 56.2% of revenue compared to (-0.9%) in the same period last year. It is also a significant improvement from (-17%) for the full year 2022 and (-6.5%) for 2023.
• Free cash flow was $46.81 million or 41.4% of revenue compared to (-$1.07 million) or (-2.9%) in the same period last year.
• Inventory was $24.4 million compared to $28.6 million in Q2.
• The company had cash and marketable securities of $886.8 million with no debt.
• Net proceeds from the IPO were $672.2 million. The shares began trading on the NASDAQ on March 20, 2024.

Valuation and IPO Risk:

On a sales valuation, Astera Labs is trading higher than Nvidia at 36X Fwd P/S compared to Nvidia’s at 29 Fwd P/S. On the bottom line, whether it’s PE Ratio, EV/EBITDA or Price to FCF, Astera is trading nearly double Nvidia’s valuation.

The company went public on March 20th, 2024 with shares opening for public trading around $62. The company raised $672.2 million with a lockup that expired September 11th, 2024. Insiders saw shares priced at $36 per share at time of IPO and the highest the stock has traded is $95 following the last earnings report.

The valuation is testing the upper limits of AI semi-related stocks. Therefore, we foresee participating now as a momentum play and participating longer-term with a new entry sometime late 2024-early 2025.

This stock requires an active stance, to where if we enter, we will exit for a quick trade, and try again at a lower valuation for a longer-term position. We offer real-time trade alerts on our Advanced tier.

Conclusion:

Management commentary is at odds with analysts’ estimates, as the commentary suggests there is a new growth phase occurring while estimates suggest a drop off in growth over the next 1-2 years. The products are enabling faster data speeds with PCIe5 and also increased back-end networking with PCIe6 for large AI clusters, and thus management’s commentary that this is a new growth phase holds weight. We certainly know the future for AI clusters is going to exponentially increase from 10s of thousands for AI clusters to eventually millions of AI accelerators per cluster. This is not only a GPU opportunity but also a custom silicon opportunity, as Astera Labs exclusively offered PCIe5 switches and retimers, and will now compete against Broadcom on PCIe6.

Astera Labs is technically in the lead and Broadcom is the follower in this case; but where it gets even more interesting is with the new product Scorpio. It is expected to increase the TAM by $5.0 billion with a total TAM of over $12B over the next 3 years. If we assume Astera captures 50% of the total TAM, then what we have is a stock that will remain in hypergrowth territory. If you do the math, that’s a potential 12X increase in revenue by 2028 from the $500M run rate Astera has today.

Certainly, technical analysis is critical given we are dealing with a very stretched valuation, a recent IPO, and ultimately, a stock that will be volatile in the years to come – yet perhaps, with the right timing, the stock will be as equally rewarding. We typically do not participate in IPOs, and Astera Labs illustrates why given its trading 2X higher than the AI juggernaut with impeccable financials (Nvidia). We do not plan to participate long-term in Astera Labs for this reason, yet may participate briefly, and then keep the stock on our watchlist from there.

Please note: The I/O Fund conducts research and draws conclusions for the Fund’s positions. We then share that information with our readers. This is not a guarantee of a stock’s performance. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis.

Recommended Reading:

• Q4 2024 Earnings Kickoff Webinar Replay
• Chainlink: A Front Runner Among Blockchain Technologies
• Cloudflare Q3 24: Soft Q4 Guide as Company Transitions on Billing Terms
• Real Vision Video Interview: Will Nvidia Continue to Dominate AI?

Astera Labs is a stock that simply cannot be ignored despite having a fraction of the market cap compared to much larger AI hardware players.

The company offers a few key products that are enabling larger and faster AI clusters. For data center AI accelerators, the company was first to offer PCIe 5 switches and retimers. Growth will likely continue due to next-generation PCIe 6 products with higher average sales prices that will be released in 2025 and ramp in 2026.

There are additional growth opportunities, including a software architecture that allows hyperscalers to monitor their data center infrastructure and increase utilization rates; the software helps to incentivize hyperscalers to use Astera Labs hardware products. The company also offers ethernet smart cable modules and is introducing a new CXL product for CXL-enabled CPUs next year.

We are interested in Astera Labs for the increased average sales prices that are expected to persist at least through 2025-2026 due to the Aries products and upcoming Scorpio products. Astera Labs is a highly technical company, yet this quote clearly communicates why the growth trajectory can sustain:

“One is generally speaking with each new generation of a protocol like PCIe going from Gen 5 to Gen 6. There is an ASP uplift. That's number one. Number two, of course, we were hinting at the Scorpio product line, which because of the value it delivers to customers is at a higher ASP, as you can imagine. So overall, if you look at the design wins we have today, the dollar content per GPU goes up, that's one way to look at it based on what we've shared before […] So overall, if you look at sort of the increasing speed, additional product line as well as the fact that the internally developed platforms, AI accelerated database platforms, they are starting to gain more and more traction. So when you look at all of them, on an average, our content is on the up.”

Investment Thesis:

• Current Growth Driver: Increased average sales prices are being driven by CPUs, GPUs and ASICs all moving to the new PCIe 5.0 standards. Arista Labs’ Aries Retimers and PCIe 5.0 components is driving the current growth, and the company is unchallenged in this new generation of PCIe, which came to market for AI accelerators only recently with Nvidia’s H200s.
• Catalyst: Scorpio is a new product that is expected to expand the TAM to more than $12B by 2028: Astera Labs is releasing a PCIe Gen 6 fabric switch custom designed for AI data flows with high performance per watt compared to incumbents. Rather than building a large switch, the company built a smaller device that is more efficient for high-speed signals. With Scorpio, Astera Labs is defending its dominance in PCIe5 by doubling the bandwidth at lower power requirements than the 5th generation of PCIe.
• The market demand for Astera Labs is healthy, as a major supplier to Nvidia’s PCIe-enabled GPUs (note: most of Nvidia’s GPUs use their in-house NVLink). Yet, Arista also supplies other AI accelerator platforms for Big Tech, and Arista is the only provider for PCIe5. In addition to hardware, ALAB offers Big Tech data centers software-defined architecture called COSMOS that allows for performance monitoring. There is some favorable vendor lock-in dynamics with COSMOS as the many different product lines can be optimized and monitored with the software.
• Financials on Astera Labs are impressive as the company is hypergrowth at 206% growth last quarter and 153% guided next quarter. The free cash flow margin is at 41%. The company is not GAAP profitable due to being a recent IPO with outsized stock-based compensation. However, there is a path to GAAP profitability, which is detailed in the financials section below.

Overview of Astera’s Products:

Broadly speaking, Astera’s products are seeing increased relevance as AI clusters grow to support hundred-billion and trillion parameter models. The company’s hardware and software increase AI server performance and productivity for the current generations and future generations of AI accelerators.

Transition to PCIe5 and PCIe5 Retimers are Driving Astera’s Growth:

The Aries products offer PCIe5 interfaces that GPUs and AI accelerators (like custom silicon) use to connect components including ethernet networking. Compared to the previous generation, PCIe5 is twice as fast with data transfer rates that reach up to 128 GBs/s on multi-directional bandwidth in each lane. By increasing the data transfer for each lane, it allows more lanes to become available to help leverage the power of the GPU or AI accelerator.

PCIe5 Retimers are chips that boost signals across high-speed components and are seeing increased demand, starting with Nvidia’s H200s, and also for application-specific chips. Specifically, Aries Retimers and smart cable modules allow hyperscalers to connect multiple racks together with up to 7 meters of copper cables. Aries can also go up to 50 meters with optical fiber. According to a presentation at Nvidia’s GTC event, this is 3X the standard reach defined in PCIe specs.

Despite PCIe5 being out since 2019, it was Nvidia’s H200s released in 2024 that were the first data center GPUs to use PCIe5. What’s interesting is that Astera is said to have captured 95% of the XPU market, which refers to application-specific chips that are specific on a product level. Per an analyst on the earnings call: “Majority of the XPU shipments are still going to be I think Gen 5 based where your market share is still somewhere in the range of, I think, like 95%.”

Management also stated: “The upside to the guidance was driven largely from Aries' revenue, both for the third-party GPUs, but also as well with the strong ramps on new platforms on the internally developed AI accelerators. And we're seeing that across multiple hyperscaler customers, so it's not just one. So the upside was largely driven by that Aries revenue.”

Notably, Aries devices are used to interconnect AI accelerators with CPUs and networking, yet are also used for backend networking between GPUs for larger clusters. Astera supplies the HGX H100 systems with PCIe-based GPUs with up to 25 retimer chips per HGX system.

Scorpio PCIe6 Custom-Built for AI is a Catalyst:

Scorpio is where the excitement is building for continued growth next year with management stating it will “exceed 10% of revenues in 2025” with “good momentum going into 2026.” This is due to the PCI Express switches being custom built for AI purposes, whereas in the past, PCIe was built for storage and then retrofitted for AI purposes.

PCIe 6 doubles the bandwidth from the 5th generation, with up to 256 GB/s of bandwidth per lane, which will require faster supporting components, such as the retimers that Astera Labs offers. The demo from GTC showed the Scorpio fabric switches (name released in October) delivering twice the bandwidth with less power at 11W instead of the 13W from the PCIe 5 interfaces.

There are two Scorpio fabric switches. The Scorpio P-Series is a small chip that connects the CPU, GPU, NIC and NVMe storage. Rather than building a large switch, the company built a smaller device that is more efficient for high-speed signals to help feed GPUs with data. The fewer ports and smaller switch decrease complexity in a bid to compete against Broadcom with twice the lane count.

The X-Series is for back-end networking in GPU-to-GPU configurations, and will offer a higher port count. Astera is essentially building something similar to Nvidia’s NVSwitch with the X-Series, but for PCIe-enabled GPUs: “And this one, like Mike noted, it's a greenfield use case, meaning if you keep Nvidia and NV Switch aside, everyone else is starting to build configurations that are obviously going to need some kind of a switching functionality, which is what we are addressing with our X Series device.” The X-Series improves efficiency for ever-increasing AI cluster sizes. The majority of AI clusters are in the tens of thousands GPUs, but are expected to go to the hundreds of thousands (already has with X and some other Big Tech companies), and will see AI clusters with millions of GPUs over the next couple of years.

Here is another quote from the management team as to why the X-Series Scorpio switch fabric is a big opportunity for their company:

“X series will have a bigger TAM. The TAM today is nearly zero. It's not very commonly used outside of the Nvidia ecosystem. We do expect many hyperscalers to start deploying this, starting with the X family and the designs for which that we have. And we are able to do that because of the architecture that we have. Because of our software-defined architecture, we can customize many parts of the X-Series to cater to the specific requirements of the hyperscalers both on the side of performance, the exact configuration that they require in count and so on and also the diagnostics framework that they require to monitor their infrastructure. So over time, we do expect X-Series to become larger [than the P-Series].”

COSMOS Software:

In order of importance, COSMOS Software ranks higher than some of Astera’s hardware as it offers performance monitoring for data center infrastructure. This is especially important for Big Tech companies concerned with utilization rates for expensive GPU systems.

The adoption of the software stack to monitor for performance is expected to increase with the Scorpio X-Series:

“Where the Scorpio family sits, we have access to a lot more diagnostic information. And we can couple that with the information that we are collecting from our other families deployed such as Aries and even Taurus to provide a holistic view of the AI infrastructure to the data center operators. So both from the hardware side, the kind of the purpose-built nature of these devices as well as the software stack that comes with it is a big differentiator for us.”

Taurus Ethernet Smart Cable Modules:

Astera Lab offers Ethernet smart cable modules which help to alleviate bandwidth issues with 100-gig per lane connectivity over copper cables including AEC. The company recently released 400-gig Ethernet SCMs, which help to stabilize the network. When thinking of ALAB’s investment thesis, Ethernet is not top-of-mind given the sheer size and lead we see from Broadcom, Nvidia’s Spectrum, and Arista Networks.

Right now, the maximum bandwidth supported by PCIe 5.0 is 400Gbps per port. By using 106Gbps PAM4 SerDes, ASICs can be tuned to support 100, 200 and 400 Gbps port speeds. To work around this, and to achieve 800Gbps, larger chip makers are building NICs directly into the accelerator. According to The Register, the 800Gbps ports built into accelerators may reduce bottlenecks before PCIe 6.0 arrives on the market. The larger Ethernet players are moving quickly on this, and we will need to keep an eye on Astera Labs to determine if the company’s Taurus product can remain competitive.

Leo Compute Express Link (CXL):

Leo is slated to impact revenue in 2025 when data center platforms plan to utilize CXL technology for memory bandwidth and capacity bottlenecks. Next year, CXL-capable CPUs will become broadly available. We’ve covered in the past how CXL is a new server architecture that “dynamically assigns memory resources between servers.” The result is boosted memory bandwidth and also at a lower cost than adding more CPUs. Partially-disaggregated racks are expected to deploy in 2024-2025 with separate compute, memory and I/O racks with the interconnect being CXL.

Per Astera’s management team: “In the past, this was done by adding additional CPUs into the server box to provide for more memory channels. But what we have demonstrated is that by using Leo you're not only able to get the higher performance by the added memory. But from an overall TCO standpoint, it's significantly less than adding additional CPUs. “

Financials: Strong Revenue Growth, Expanding Margins

Astera Labs reported solid top-line and bottom-line beats in the recent Q3 results. The company reported record revenue of $113.1 million, up 47% QoQ and 206% YoY, beating estimates by 16.1%. The adjusted operating margin expanded to 32.4% from 2% in the same period last year, which was helped by strong operating leverage. The adjusted EPS of $0.23 beat estimates by 35%.

Revenue:

The company is one of the fastest-growing tech companies and is emerging as a rising star in the AI data center networking space. The company’s Q3 revenue grew by 206.2% YoY and 47% QoQ to $113.1 million.

Jitendra Mohan, CEO and co-founder of the company, said in the Q3 2024 earnings call, “Our business has entered a new growth phase with multiple product families ramping across AI platforms, featuring both third-party GPUs and internally developed AI accelerators, which drove the Q3 sales upside verus our guidance.”

• Management also provided a strong Q4 revenue guide of $126 million to $130 million, representing YoY growth of 153.4% at the mid-point. This also represents QoQ growth of 13% at the midpoint.

  According to management, the QoQ growth is being driven by “our Aries product family across a diverse set of AI platforms, some of which are just starting to ramp and also from our Taurus SCM for 400-gig applications, and additional preproduction shipments of our Scorpio P-Series switches.”

• Analysts estimate revenue to grow 103.5% to $132.78 million in Q1 2025 and 85.2% YoY to $142.34 million in Q2 2025. While the growth rates are strong, growth is expected to slow down due to tougher comps.

Jitendra Mohan said, “Looking into Q4, we expect our revenue momentum to continue, largely driven by the Aries PCIe and Taurus Ethernet product lines. The Scorpio Fabric Switches are continuing to ship in preproduction volumes.”

The newly introduced Scorpio Fabric Switches are expected to increase the total market opportunity to more than $12 billion by 2028 for the company. Scorpio Switches are also expected to constitute more than 10% of revenue in 2025.

Source: Company website

Analysts expect 2024 revenue growth of 230.9% YoY to $383.14 million, followed by 55.5% and 40.4% in the subsequent years. Meanwhile, management comments seem to imply the growth recently reported will sustain, implying the growth phase has only begun: “Our business has entered a new growth phase with multiple product families ramping across AI platforms, featuring both third-party GPUs and internally developed AI accelerators, which drove the Q3 sales upside versus our guidance.”

Management also later stated: “Yes, right now, our visibility is very strong, both as always with our backlog position, but also the breadth of designs we have — right now, we're really kind of entering a new phase of growth here where our revenue streams are clearly diversifying […].”

Margins:

The company’s margins are improving, helped by strong operating leverage. However, the product mix might weigh on the margins going forward. Management mentioned in the recent earnings call that they have a long-term gross margin target of 70%.

• Q3 gross profits grew by 212.7% YoY to $87.88 million or 77.7% of revenue compared to 76.1% in the same period last year.
• This compares to 73.5% for FY2022 and 68.9% for FY2023.
• Adjusted gross margin was 77.8% compared to 76.1% in the same period last year.
• Management gross margin and adjusted gross margin guide for the next quarter are 75%. This is down sequentially due to higher product mix towards hardware solutions during the quarter.
• Operating margin was (-7.9%) compared to (-5.3%) in the same period last year. The adjusted operating margin expanded to 32.4% from 2% in the same period last year, helped by strong operating leverage.
• Management operating margin guide for the next quarter is (-4.3%) and adjusted operating margin guide is 32.4%.
• It’s important to consider that stock-based compensation is quite high due to the recent IPO at 40.3% this quarter, and was at 56% last quarter. Once SBC naturally levels out, this company has strong enough margins to become GAAP profitable.
• Net loss was (-$7.6 million) or (-6.7%) of revenue compared to (-8.5%) in the same period last year.
• Adjusted net income improved significantly to $40.28 million or 35.6% of revenue compared to (-$0.41 million) or (-1.1%) of revenue in the same period last year.

The difference between the GAAP and non-GAAP net income is due to high stock-based compensation. Stock-based compensation was $45.5 million or 40.3% of revenue in the recent quarter. Stock-based compensation has been lumpy as the company’s IPO was in March 2024.

EPS:

The company beat EPS estimates helped by solid operating leverage. Q3 GAAP loss per share was ($0.05) compared to ($0.08) in the same period last year, beating estimates by 28.1%. Adjusted EPS was $0.23 and beat estimates by 35%.

• Management has guided the Q4 GAAP EPS in the range of $0.04 to $0.06.
• Notably, the company will not be GAAP operating income positive next quarter as the guide suggests an operating loss of (-$5.5 million) and interest income of about $10 million in the next quarter will make it GAAP profitable. However, it’s very close to being GAAP profitable across all margins, and we think it’s only a matter of time before this happens.
• Adjusted EPS guide is $0.25 to $0.26.
• Analysts expect adjusted EPS to be $0.25 in Q1 2025 and $0.27 in Q2 2025.
• Analysts expect strong EPS growth. They expect 2025 adjusted EPS to grow 58.9% YoY to $1.14 and 48.6% YoY to $1.70 in 2026.

Cash Flow and Balance Sheet:

The company’s cash flows margins are high and are also improving due to increased profits.

• Q3 operating cash flow was $63.5 million or 56.2% of revenue compared to (-0.9%) in the same period last year. It is also a significant improvement from (-17%) for the full year 2022 and (-6.5%) for 2023.
• Free cash flow was $46.81 million or 41.4% of revenue compared to (-$1.07 million) or (-2.9%) in the same period last year.
• Inventory was $24.4 million compared to $28.6 million in Q2.
• The company had cash and marketable securities of $886.8 million with no debt.
• Net proceeds from the IPO were $672.2 million. The shares began trading on the NASDAQ on March 20, 2024.

Valuation and IPO Risk:

On a sales valuation, Astera Labs is trading higher than Nvidia at 36X Fwd P/S compared to Nvidia’s at 29 Fwd P/S. On the bottom line, whether it’s PE Ratio, EV/EBITDA or Price to FCF, Astera is trading nearly double Nvidia’s valuation.

The company went public on March 20th, 2024 with shares opening for public trading around $62. The company raised $672.2 million with a lockup that expired September 11th, 2024. Insiders saw shares priced at $36 per share at time of IPO and the highest the stock has traded is $95 following the last earnings report.

The valuation is testing the upper limits of AI semi-related stocks. Therefore, we foresee participating now as a momentum play and participating longer-term with a new entry sometime late 2024-early 2025.

This stock requires an active stance, to where if we enter, we will exit for a quick trade, and try again at a lower valuation for a longer-term position. We offer real-time trade alerts on our Advanced tier.

Conclusion:

Management commentary is at odds with analysts’ estimates, as the commentary suggests there is a new growth phase occurring while estimates suggest a drop off in growth over the next 1-2 years. The products are enabling faster data speeds with PCIe5 and also increased back-end networking with PCIe6 for large AI clusters, and thus management’s commentary that this is a new growth phase holds weight. We certainly know the future for AI clusters is going to exponentially increase from 10s of thousands for AI clusters to eventually millions of AI accelerators per cluster. This is not only a GPU opportunity but also a custom silicon opportunity, as Astera Labs exclusively offered PCIe5 switches and retimers, and will now compete against Broadcom on PCIe6.

Astera Labs is technically in the lead and Broadcom is the follower in this case; but where it gets even more interesting is with the new product Scorpio. It is expected to increase the TAM by $5.0 billion with a total TAM of over $12B over the next 3 years. If we assume Astera captures 50% of the total TAM, then what we have is a stock that will remain in hypergrowth territory. If you do the math, that’s a potential 12X increase in revenue by 2028 from the $500M run rate Astera has today.

Certainly, technical analysis is critical given we are dealing with a very stretched valuation, a recent IPO, and ultimately, a stock that will be volatile in the years to come – yet perhaps, with the right timing, the stock will be as equally rewarding. We typically do not participate in IPOs, and Astera Labs illustrates why given its trading 2X higher than the AI juggernaut with impeccable financials (Nvidia). We do not plan to participate long-term in Astera Labs for this reason, yet may participate briefly, and then keep the stock on our watchlist from there.

Recommended Reading:

• Q4 2024 Earnings Kickoff Webinar Replay
• Chainlink: A Front Runner Among Blockchain Technologies
• Cloudflare Q3 24: Soft Q4 Guide as Company Transitions on Billing Terms
• Real Vision Video Interview: Will Nvidia Continue to Dominate AI?

AppLovin has done the unthinkable, which is to awaken a low-growth mobile gaming ads industry with an AI engine that is showing demonstrable results. The market is loving this stock as it has doubled its margins, more than doubled its cash flow and has a surging AI segment due to its AXON 2.0 AI advertising engine.

Our near-term plan is to trade this stock, while our medium-term plan is to build a longer-term position. Both require an active stance rather than guessing on the buys. However, we think App is setting up for a longer-term trajectory and our firm plans to participate.

Update on Investment Thesis:

There are a few key points to the investment thesis that I’d like to bookmark here for future reference.

APP has User Data from 1.4 Billion Mobile Users

The first point to the longer-term investment thesis is that AppLovin has data from 1.4 billion mobile gamers. We’ve seen the razor-razor blade model with hardware, to where a company will own the hardware market to get recurring software revenue. AppLovin has a different variation of this, which is they own mobile gaming apps and a supply-side platform to mix both first-party data and third-party data, which in turn, fuels their AI engine to help them capitalize on the broader mobile gaming market.

The word moat is overused in tech stocks, yet AppLovin has an enviable advantage in the era of AI. Off the top of my head, I cannot think of another company with this level of user data for advertising purposes that is not a Big Tech company and in the Mag 7.

Catalyst: E-commerce and more Web-based Advertising Categories in 2025

AppLovin’s success has been entirely based on gaming companies advertising to mobile gamers.

The company is planning to introduce new advertising segments to the 1.4 billion users they serve in 2025, which is likely to help the company grow into the foreseeable future. The catalyst for 2025 is expected to broaden to also include a self-service platform for all types of web-based advertising. According to management, the pilot for introducing e-commerce demand is going quite well: “E-commerce, on the other hand, is looking so strong that it's something that we think will be impactful to the business financially '25 and then for the long term.”

Product Differentiation:

The ad engine AXON 2.0 offers a monumental advantage to AppLovin as the company is ahead in the race for AI-driven advertising. Part of this is the user data from 1.4 billion users, which cannot be overemphasized, and it’s also due to the company owning both a supply-side platform MAX and demand-side platform, App Discovery. MAX gives AppLovin data on what different ad networks are willing to bid for ad placements, allowing AXON to competitively bid for ad placements to maximize return-on-ad-spend. After Apple announced its App Tracking Transparency (ATT) policy in 2021 which limited advertisers’ ability to track users across apps, AppLovin’s data became even more valuable as advertisers sought out AppDiscovery’s user acquisition algorithms to acquire high value users cost-effectively.

AppLovin is also an arbitrage advertising platform, which means they can quantify the impact of their reach for advertisers by returning back to the advertiser what was spent or more within 30 days. If an advertiser spends $10,000 (or multiples of this), AppLovin is able to return that or more to the advertiser. The company is also unique in that it offers performance marketing for brands and direct-to-consumer. The Trade Desk primarily works with agencies, whereas AppLovin is attracting smaller and medium sized businesses that rely on performance.

Most importantly, AXON 2.0 is an AI-powered advertising engine that is continuously improving. Every quarter and every year, AXON becomes more effective by ingesting more data that improves the model through self-learning. The management has been quite clear they believe these step-ups in model efficiency can help to maintain a 20% to 30% growth rate in gaming alone, and not accounting for the new web-based advertising catalysts expected in 2025.

“Last quarter, I shared our confidence in achieving 20% to 30% year-over-year growth for the foreseeable future. We continue to expect 4% to 5% quarterly growth through self-learning and market growth, with occasional step changes resulting from enhancements to our AXON algorithm.”

Strong Bottom Line and Cash Flows:

On top of the growth potential, APP has a strong bottom line and cash flows, which we review in more detail below. The adjusted EBITDA margin is at 60% and the GAAP operating margin has doubled YoY from 21.6% to 44.6%. The free cash flow margin of 45.5% has also doubled from 22.4%. Notably, the company carries $3.5 billion in debt, yet at this cash flow margin is not a concern.

You can read more about AppLovin here.about AppLovin here.

Q3 Earnings: Software/Advertising Segment up 16% QoQ

In addition to reporting growth on the bottom line, AppLovin’s primary AI segment inflected 17% QoQ to $835 million, up from $711 million last quarter. The software/advertising platform has a high adjusted EBITDA of 78%, easily making this one of the more profitable hypergrowth companies the market has ever seen. This is not exactly a secret, as AppLovin is up over 600% YTD but what is important to look at it, is whether AppLovin can continue this winning streak.

Given commentary on the most recent earnings call, we think analyst estimates are too low for next year. There’s also a valuation case being made by institutional analysts following the last earnings report that APP should be valued by EV/EBITDA, which creates room in the valuation that traditional top line and bottom-line metrics are not showing.

Revenue

APP reported $1.2 billion in revenue in Q3, beating estimates by nearly 6% after missing slightly in Q2. Revenue growth continued to decelerate from its peak of 47.9% YoY in Q1, with Q3 revenue growth of 38.6% YoY.

For Q4, APP guided for revenue between $1.24 to $1.26 billion, or 31.1% YoY growth at the midpoint, pointing to growth decelerating once more as comps get tougher. Moving through the first half of 2025, growth is expected to hover in the low-20% range, up from the mid-teens before the report.

The reason the market is ignoring the deceleration is that the key AI segment is growing QoQ and re-accelerated in the most recent quarter. This hints at a re-acceleration potentially in the top line in the coming quarters.

For FY24, APP is expected to see revenue rise 39.9% YoY to $4.59 billion, before slowing to 18.9% YoY growth to $5.46 billion in FY25.

Our firm is tracking the 2025 catalysts of e-commerce and other web-based advertising segments as offering strong potential that analyst estimates are too low, especially when coupled with management comments that gaming alone will drive 20% to 30% revenue growth into the foreseeable future.

Margins:

APP’s margin strengths have been an underlying driver of the surge in the stock price, with increased operating leverage driving a strong expansion on the bottom line.

• Gross margin in Q3 was 77.5%, improving from 73.8% in Q2 and 69.3% a year ago.
• Operating margin in Q3 was 44.6%, a significant improvement from 36.2% in Q2 and more than double the 21.6% operating margin in the year ago quarter. This degree of operating leverage is ridiculous! Especially while seeing revenue growth rates above 30% — essentially, APP has been able to drive this revenue growth with barely any change to its operating expenses, even as it continues to improves its AXON AI engine. Wow.
• Net margin in Q3 was 36.3%, improving from 28.7% in Q2 and nearly triple the 12.6% margin in the year ago quarter, due to that substantial operating leverage. Again, wow.
• Adjusted EBITDA margin was 60% in Q3, up from 56% in Q2 and 49% in the year ago quarter. For Q4, APP guided adjusted EBITDA margin to remain flat QoQ at 60%.

EPS

Given the dramatic improvement in operating and net margins, APP’s net income and EPS has followed suit, rising over 300% YoY in Q3.

Q3’s GAAP EPS of $1.25 increased 317% YoY, and easily beat estimates for $0.93. This accelerated slightly from Q2’s 304% YoY growth. Looking ahead, GAAP EPS is expected to remain flat QoQ in Q4 at $1.25, before advancing slightly in the first half of 2025 to the mid-$1.30 range.

Through Q3, APP’s GAAP EPS has risen 462% YoY to $2.81. Using Q4’s guide, FY24’s EPS would be estimated at $4.06, or YoY growth of 314%.

For FY25 and FY26, EPS growth is expected to remain robust even after this surge, with growth projected currently at >30% in both years: analysts estimate 37% YoY growth to $5.57 in FY25, and 31% YoY growth to $7.26 in FY26.

Cash and Balance Sheet:

Operating cash flow and free cash flow growth has also been remarkably strong, with margins quickly approaching 50%.

• Operating cash flow was $550.7 million in Q3, increasing 177% YoY. OCF margin was 46%, improving from 42.1% in Q2 and doubling from 23% in the year ago quarter.
• Free cash flow was $545.1 million in Q3, rising 182% YoY. FCF margin was 45.5%, improving from 41.2% last quarter and 22.4% in the year ago quarter.
• Cash and equivalents totaled $567.6 million.
• Debt totaled $3.51 billion.

What’s of note here is that APP carries a high debt load, with the first $1.5 billion tranche of senior secured term loans due in 2028, with the remaining $2.1 billion senior secured loan due in 2030.

Key Segments and Metrics

APP’s Software segment (soon to be reclassified as Advertising) and its AXON AI engine has been the primary growth driver over the course of the past six quarters, with Software Platform revenue rising from 50% of total revenue at the beginning of 2023 to 70% in Q3 2024.

Software’s growth remained strong in Q3, with revenue rising 66% YoY to $835 million. This marked the fifth straight quarter of YoY growth >60% for the segment, with growth also reaccelerating sequentially, with QoQ of 17.4% in Q3 versus 4.8% in Q2.

Software’s adjusted EBITDA increased 79% YoY to $653 million, outpacing revenue growth as a result of increased operating leverage. Adjusted EBITDA margin in the segment was 78%, expanding from 73% last quarter and 72% in the year ago quarter.

On the other hand, App revenue was relatively unchanged, rising just 1% YoY to $363 million, and decelerating from 7% growth in the prior quarter. App’s revenue has not yet rebounded after a trough in early 2023.

App’s adjusted EBITDA rose nearly 24% YoY to $68 million, or a 19% margin. This contracted slightly from 22% in Q2 but had improved from 15% in the year ago quarter; however, segment performance remains slightly challenged as APP continues to optimize the segment’s cost structure.

Monthly Active Payers, Average Revenue:

APP’s monthly active payers were 1.6 million, flat sequentially but down from 1.8 million last year. On the other hand, average revenue per monthly active payer (ARPMAP) was $52, flat sequentially but improving from $46 last year.

Earnings Call Discussion:

Bull Case: 20% to 30%+ Growth into Next Year

As stated, the comments that gaming alone can drive 20% to 30% growth, in addition to the important catalyst of expanding into e-commerce and web-based advertising is why AppLovin can continue on a strong growth trajectory. Analysts currently have FY2025 estimates at 19.6%:

Here is the tone from the earnings call:

“While we remain confident in 20% to 30% growth for mobile gaming advertisers alone, we're also exploring new areas, as shown by our recent e-commerce pilot. Early data has exceeded our expectations, with the advertisers in the pilot seeing substantial returns, often surpassing those from other media channels, and in many cases, experiencing nearly a 100% incrementality from our traffic.”

The Scale + AI Engine is Why AppLovin is Just Getting Started

AppLovin’s 1.4 billion users is key to why this company’s trajectory may just be getting started. The company has stated even if they release the code and algorithm for AXON 2.0, competitors cannot mimic what they’ve built due to the data they own.

Here is what was stated on the call:

“[Our customers] care about optimization and automated advertising to a revenue goal. And that's really like what our system is predicated on is that. We take all the risk on the media side. We have to deliver really compelling performance on the technology side. And I guess, like what's most exciting for me on what we've built and where we are in terms of, like you said, market cap and scale as a business today is we're on top of 1.4 billion daily actives. So it's really easy to forget the scale of the audience reach that we have on our platform. We've got the largest mediation solution in the sector, and our teams built maybe the most innovative advertising technology that the world has yet seen.”

Valuation:

EV/EBITDA Valuation Shows Room

No doubt, AppLovin is richly valued on the top line and bottom line. The top line is trading at 21X Fwd PS compared to ad-tech peer The Trade Desk at 26X Fwd PS. On the bottom line, AppLovin is trading at 52 Fwd PE Ratio compared to a Fwd PE Ratio of 79 for TTD.

However, where there is more room is seen in the EV/EBITDA valuation, which institutional analysts are making the case should be the correct valuation. Per current data, there is nearly 100% upside to APP on this valuation.

Here is what analysts are saying: “(11/07) Macquarie raised the firm's price target on AppLovin to $270 from $150 and keeps an Outperform rating on the shares after the company's beat and raise Q3 report. The firm, which raised its 2024 adjusted EBITDA estimate to $2.6B from $2.4B and its 2025 estimate to $3.2B from $2.9B, notes that it shifted its valuation method to "straight EV/EBITDA, given the predominance of the Software Platform now."

Source: YCharts

Conclusion:

Our firm is preparing to participate in AppLovin in two ways — first, a quicker momentum play to see if we can capture any remaining upside presented in the EV/EBITDA valuation that institutional analysts are favoring for this stock. Knox’s technical analysis is showing a potential move, and this is supported by fundamentals. However, for our longer-term position, we will look to close the momentum trade and participate again come 2025. Stay tuned!

Please note: The I/O Fund conducts research and draws conclusions for the Fund’s positions. We then share that information with our readers. This is not a guarantee of a stock’s performance. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis.

Recommended Reading:

• AOSL Q1 2025: Foreshadowing Consumer Weakness
• Coinbase Q3 24: Strong Cash and Adjusted EBITDA; Technicals Matter
• Nova and Onto Innovation: Growth in Metrology and Semiconductor Process Control
• AppLovin Corporation: Emerging Ad Tech AI Leader

AppLovin has done the unthinkable, which is to awaken a low-growth mobile gaming ads industry with an AI engine that is showing demonstrable results. The market is loving this stock as it has doubled its margins, more than doubled its cash flow and has a surging AI segment due to its AXON 2.0 AI advertising engine.

Our near-term plan is to trade this stock, while our medium-term plan is to build a longer-term position. Both require an active stance rather than guessing on the buys. However, we think App is setting up for a longer-term trajectory and our firm plans to participate.

Update on Investment Thesis:

There are a few key points to the investment thesis that I’d like to bookmark here for future reference.

APP has User Data from 1.4 Billion Mobile Users

The first point to the longer-term investment thesis is that AppLovin has data from 1.4 billion mobile gamers. We’ve seen the razor-razor blade model with hardware, to where a company will own the hardware market to get recurring software revenue. AppLovin has a different variation of this, which is they own mobile gaming apps and a supply-side platform to mix both first-party data and third-party data, which in turn, fuels their AI engine to help them capitalize on the broader mobile gaming market.

The word moat is overused in tech stocks, yet AppLovin has an enviable advantage in the era of AI. Off the top of my head, I cannot think of another company with this level of user data for advertising purposes that is not a Big Tech company and in the Mag 7.

Catalyst: E-commerce and more Web-based Advertising Categories in 2025

AppLovin’s success has been entirely based on gaming companies advertising to mobile gamers.

The company is planning to introduce new advertising segments to the 1.4 billion users they serve in 2025, which is likely to help the company grow into the foreseeable future. The catalyst for 2025 is expected to broaden to also include a self-service platform for all types of web-based advertising. According to management, the pilot for introducing e-commerce demand is going quite well: “E-commerce, on the other hand, is looking so strong that it's something that we think will be impactful to the business financially '25 and then for the long term.”

Product Differentiation:

The ad engine AXON 2.0 offers a monumental advantage to AppLovin as the company is ahead in the race for AI-driven advertising. Part of this is the user data from 1.4 billion users, which cannot be overemphasized, and it’s also due to the company owning both a supply-side platform MAX and demand-side platform, App Discovery. MAX gives AppLovin data on what different ad networks are willing to bid for ad placements, allowing AXON to competitively bid for ad placements to maximize return-on-ad-spend. After Apple announced its App Tracking Transparency (ATT) policy in 2021 which limited advertisers’ ability to track users across apps, AppLovin’s data became even more valuable as advertisers sought out AppDiscovery’s user acquisition algorithms to acquire high value users cost-effectively.

AppLovin is also an arbitrage advertising platform, which means they can quantify the impact of their reach for advertisers by returning back to the advertiser what was spent or more within 30 days. If an advertiser spends $10,000 (or multiples of this), AppLovin is able to return that or more to the advertiser. The company is also unique in that it offers performance marketing for brands and direct-to-consumer. The Trade Desk primarily works with agencies, whereas AppLovin is attracting smaller and medium sized businesses that rely on performance.

Most importantly, AXON 2.0 is an AI-powered advertising engine that is continuously improving. Every quarter and every year, AXON becomes more effective by ingesting more data that improves the model through self-learning. The management has been quite clear they believe these step-ups in model efficiency can help to maintain a 20% to 30% growth rate in gaming alone, and not accounting for the new web-based advertising catalysts expected in 2025.

“Last quarter, I shared our confidence in achieving 20% to 30% year-over-year growth for the foreseeable future. We continue to expect 4% to 5% quarterly growth through self-learning and market growth, with occasional step changes resulting from enhancements to our AXON algorithm.”

Strong Bottom Line and Cash Flows:

On top of the growth potential, APP has a strong bottom line and cash flows, which we review in more detail below. The adjusted EBITDA margin is at 60% and the GAAP operating margin has doubled YoY from 21.6% to 44.6%. The free cash flow margin of 45.5% has also doubled from 22.4%. Notably, the company carries $3.5 billion in debt, yet at this cash flow margin is not a concern.

You can read more about AppLovin here.about AppLovin here.

Q3 Earnings: Software/Advertising Segment up 16% QoQ

In addition to reporting growth on the bottom line, AppLovin’s primary AI segment inflected 17% QoQ to $835 million, up from $711 million last quarter. The software/advertising platform has a high adjusted EBITDA of 78%, easily making this one of the more profitable hypergrowth companies the market has ever seen. This is not exactly a secret, as AppLovin is up over 600% YTD but what is important to look at it, is whether AppLovin can continue this winning streak.

Given commentary on the most recent earnings call, we think analyst estimates are too low for next year. There’s also a valuation case being made by institutional analysts following the last earnings report that APP should be valued by EV/EBITDA, which creates room in the valuation that traditional top line and bottom-line metrics are not showing.

Revenue

APP reported $1.2 billion in revenue in Q3, beating estimates by nearly 6% after missing slightly in Q2. Revenue growth continued to decelerate from its peak of 47.9% YoY in Q1, with Q3 revenue growth of 38.6% YoY.

For Q4, APP guided for revenue between $1.24 to $1.26 billion, or 31.1% YoY growth at the midpoint, pointing to growth decelerating once more as comps get tougher. Moving through the first half of 2025, growth is expected to hover in the low-20% range, up from the mid-teens before the report.

The reason the market is ignoring the deceleration is that the key AI segment is growing QoQ and re-accelerated in the most recent quarter. This hints at a re-acceleration potentially in the top line in the coming quarters.

For FY24, APP is expected to see revenue rise 39.9% YoY to $4.59 billion, before slowing to 18.9% YoY growth to $5.46 billion in FY25.

Our firm is tracking the 2025 catalysts of e-commerce and other web-based advertising segments as offering strong potential that analyst estimates are too low, especially when coupled with management comments that gaming alone will drive 20% to 30% revenue growth into the foreseeable future.

Margins:

APP’s margin strengths have been an underlying driver of the surge in the stock price, with increased operating leverage driving a strong expansion on the bottom line.

• Gross margin in Q3 was 77.5%, improving from 73.8% in Q2 and 69.3% a year ago.
• Operating margin in Q3 was 44.6%, a significant improvement from 36.2% in Q2 and more than double the 21.6% operating margin in the year ago quarter. This degree of operating leverage is ridiculous! Especially while seeing revenue growth rates above 30% — essentially, APP has been able to drive this revenue growth with barely any change to its operating expenses, even as it continues to improves its AXON AI engine. Wow.
• Net margin in Q3 was 36.3%, improving from 28.7% in Q2 and nearly triple the 12.6% margin in the year ago quarter, due to that substantial operating leverage. Again, wow.
• Adjusted EBITDA margin was 60% in Q3, up from 56% in Q2 and 49% in the year ago quarter. For Q4, APP guided adjusted EBITDA margin to remain flat QoQ at 60%.

EPS

Given the dramatic improvement in operating and net margins, APP’s net income and EPS has followed suit, rising over 300% YoY in Q3.

Q3’s GAAP EPS of $1.25 increased 317% YoY, and easily beat estimates for $0.93. This accelerated slightly from Q2’s 304% YoY growth. Looking ahead, GAAP EPS is expected to remain flat QoQ in Q4 at $1.25, before advancing slightly in the first half of 2025 to the mid-$1.30 range.

Through Q3, APP’s GAAP EPS has risen 462% YoY to $2.81. Using Q4’s guide, FY24’s EPS would be estimated at $4.06, or YoY growth of 314%.

For FY25 and FY26, EPS growth is expected to remain robust even after this surge, with growth projected currently at >30% in both years: analysts estimate 37% YoY growth to $5.57 in FY25, and 31% YoY growth to $7.26 in FY26.

Cash and Balance Sheet:

Operating cash flow and free cash flow growth has also been remarkably strong, with margins quickly approaching 50%.

• Operating cash flow was $550.7 million in Q3, increasing 177% YoY. OCF margin was 46%, improving from 42.1% in Q2 and doubling from 23% in the year ago quarter.
• Free cash flow was $545.1 million in Q3, rising 182% YoY. FCF margin was 45.5%, improving from 41.2% last quarter and 22.4% in the year ago quarter.
• Cash and equivalents totaled $567.6 million.
• Debt totaled $3.51 billion.

What’s of note here is that APP carries a high debt load, with the first $1.5 billion tranche of senior secured term loans due in 2028, with the remaining $2.1 billion senior secured loan due in 2030.

Key Segments and Metrics

APP’s Software segment (soon to be reclassified as Advertising) and its AXON AI engine has been the primary growth driver over the course of the past six quarters, with Software Platform revenue rising from 50% of total revenue at the beginning of 2023 to 70% in Q3 2024.

Software’s growth remained strong in Q3, with revenue rising 66% YoY to $835 million. This marked the fifth straight quarter of YoY growth >60% for the segment, with growth also reaccelerating sequentially, with QoQ of 17.4% in Q3 versus 4.8% in Q2.

Software’s adjusted EBITDA increased 79% YoY to $653 million, outpacing revenue growth as a result of increased operating leverage. Adjusted EBITDA margin in the segment was 78%, expanding from 73% last quarter and 72% in the year ago quarter.

On the other hand, App revenue was relatively unchanged, rising just 1% YoY to $363 million, and decelerating from 7% growth in the prior quarter. App’s revenue has not yet rebounded after a trough in early 2023.

App’s adjusted EBITDA rose nearly 24% YoY to $68 million, or a 19% margin. This contracted slightly from 22% in Q2 but had improved from 15% in the year ago quarter; however, segment performance remains slightly challenged as APP continues to optimize the segment’s cost structure.

Monthly Active Payers, Average Revenue:

APP’s monthly active payers were 1.6 million, flat sequentially but down from 1.8 million last year. On the other hand, average revenue per monthly active payer (ARPMAP) was $52, flat sequentially but improving from $46 last year.

Earnings Call Discussion:

Bull Case: 20% to 30%+ Growth into Next Year

As stated, the comments that gaming alone can drive 20% to 30% growth, in addition to the important catalyst of expanding into e-commerce and web-based advertising is why AppLovin can continue on a strong growth trajectory. Analysts currently have FY2025 estimates at 19.6%:

Here is the tone from the earnings call:

“While we remain confident in 20% to 30% growth for mobile gaming advertisers alone, we're also exploring new areas, as shown by our recent e-commerce pilot. Early data has exceeded our expectations, with the advertisers in the pilot seeing substantial returns, often surpassing those from other media channels, and in many cases, experiencing nearly a 100% incrementality from our traffic.”

The Scale + AI Engine is Why AppLovin is Just Getting Started

AppLovin’s 1.4 billion users is key to why this company’s trajectory may just be getting started. The company has stated even if they release the code and algorithm for AXON 2.0, competitors cannot mimic what they’ve built due to the data they own.

Here is what was stated on the call:

“[Our customers] care about optimization and automated advertising to a revenue goal. And that's really like what our system is predicated on is that. We take all the risk on the media side. We have to deliver really compelling performance on the technology side. And I guess, like what's most exciting for me on what we've built and where we are in terms of, like you said, market cap and scale as a business today is we're on top of 1.4 billion daily actives. So it's really easy to forget the scale of the audience reach that we have on our platform. We've got the largest mediation solution in the sector, and our teams built maybe the most innovative advertising technology that the world has yet seen.”

Valuation:

EV/EBITDA Valuation Shows Room

No doubt, AppLovin is richly valued on the top line and bottom line. The top line is trading at 21X Fwd PS compared to ad-tech peer The Trade Desk at 26X Fwd PS. On the bottom line, AppLovin is trading at 52 Fwd PE Ratio compared to a Fwd PE Ratio of 79 for TTD.

However, where there is more room is seen in the EV/EBITDA valuation, which institutional analysts are making the case should be the correct valuation. Per current data, there is nearly 100% upside to APP on this valuation.

Here is what analysts are saying: “(11/07) Macquarie raised the firm's price target on AppLovin to $270 from $150 and keeps an Outperform rating on the shares after the company's beat and raise Q3 report. The firm, which raised its 2024 adjusted EBITDA estimate to $2.6B from $2.4B and its 2025 estimate to $3.2B from $2.9B, notes that it shifted its valuation method to "straight EV/EBITDA, given the predominance of the Software Platform now."

Source: YCharts

Conclusion:

Our firm is preparing to participate in AppLovin in two ways — first, a quicker momentum play to see if we can capture any remaining upside presented in the EV/EBITDA valuation that institutional analysts are favoring for this stock. Knox’s technical analysis is showing a potential move, and this is supported by fundamentals. However, for our longer-term position, we will look to close the momentum trade and participate again come 2025. Stay tuned!

Recommended Reading:

• AOSL Q1 2025: Foreshadowing Consumer Weakness
• Coinbase Q3 24: Strong Cash and Adjusted EBITDA; Technicals Matter
• Nova and Onto Innovation: Growth in Metrology and Semiconductor Process Control
• AppLovin Corporation: Emerging Ad Tech AI Leader

Chainlink is one of our favorite long-term plays for a ten-year holding period or more. The number of stocks our firm has held without interruption since the day we launched our site in July of 2019 is very few, not even Microsoft makes that list. But, Chainlink does.

The company is a reliable source for off-chain data across all blockchain networks and blockchain apps. Chainlink’s interoperability causes its oracle networks to be leveraged and trusted across competing blockchains. Customers that require the highest level of security, such as bank transfers, rely on Chainlink’s oracle price feeds and the other off-chain data that Chainlink supplies.

To understand the problem that Chainlink solves, it’s important to discuss the overarching issues the blockchain faces. There are three necessary tenets to a fully developed blockchain protocol: security, decentralization and scalability. This trifecta is often called the “trilemma” as blockchain technologies can typically solve for two of these problems, yet struggle to solve all three. Instead, current blockchain technologies excel at two of the three, and then must experiment to solve the third with sidechains, sharding and other nascent attempts at solving the third requirement.

The analysis below discusses how Chainlink solves some of the toughest pain points in blockchain development by enabling the third, missing pillar – whether that be security, decentralization or scalability. Because Chainlink is both middleware and a Layer 2 technology, its oracle networks are utilized across nearly all leading blockchain networks plus hundreds of applications. By offering must-have solutions early-on, Chainlink is proving it can become a frontrunner for the day when blockchain reaches critical mass.

I realize the information below can be technical, and thus I’d like to provide a 10,000 foot overview of what is being described in the deep dive.

Summary:Summary: Chainlink has the key ingredients to become a front runner in blockchain technologies because:

• it can serve nearly all Layer 1s and Layer 2s for the ultimate addressable market
• Chainlink’s suite of technologies solve some of the hardest problems that blockchain technologies faces today 
• An oracle network for offchain data can form a moat — as the quality of its security and decentralization increases– it will be harder to disrupt as time goes on. Blockchain technologies will not want to take a chance on a new entrant ten years from now. I foresee a duopoly of sorts for oracle networks.
• The finance sector has essentially adopted two blockchain technologies: Bitcoin as a store of value, and Chainlink for its ability to facilitate transfers with offchain data. Given there have been thousands of blockchain technologies, it’s clear to see front runners are being adopted by the savvy $6.4 trillion banking sector.

Section 1: How Chainlink Solves Blockchain’s Trilemma

Bitcoin’s network has prioritized security and decentralization, while sacrificing scalability. The network sends 7 transactions per second and can take up to 10 minutes to confirm a transaction. The upside is the network’s security is bullet-proof with a hash rate of 460 Exahash per second. It’s impossible today for a super computer to crack the Proof-of-Work encryption.

When you add the fact there are thousands of nodes globally, a Proof of Work system is truly decentralized whereas a Proof of Stake (PoS) system could still concentrate itself through “whales,” those who own a disproportionate amount of a single token. This results in the wealthiest crypto holders having a higher concentration in what is essentially a lottery system of validators. If a person has a thousand lottery tickets compared to a person with only ten, the person with 1,000 tickets (or nodes in this case) is more likely to be chosen to validate the ledger. This could lead to corruption and ultimately does not fit crypto’s ethos that those with a higher concentration of wealth are allowed to be more trusted and have more authority.

It’s been generally understood for some time that Ethereum has prioritized security and decentralization over scalability. Despite scalability being the primary problem Ethereum has left to solve, there are solid debates that Ethereum is not as secure or decentralized as it once was following the merge to Proof of Stake (PoS). To understand these concerns, consider that Ethereum has seen up to 70% of its supply held by whales in 2021, although the latest report is that 43% of ETH supply is held by whales. The concentration is staggering as six of the top crypto wallets have 98% of their wallets allocated to the Ethereum blockchain, according to TechCrunch.

When looking more closely at Proof of Stake (PoS) validators for Ethereum, Lido is the largest Ethereum validator at 33% stake and Coinbase is at 15%. To help illustrate how unusual this concentration is, consider that the Nakamoto Coefficient for Ethereum is 2, which means it takes only two nodes to control the blockchain. Truly, it’s beyond belief the coefficient is this low for the world’s top blockchain Layer 1. Bitcoin’s is estimated to be as high as 9601. The highest Coefficient beyond Bitcoin is 236 for a network called Humanode, and its goal is to increase the coefficient over time. The last time Solana’s was reported in 2023, it stood at a coefficient of 31.

Also consider that PoS requires 32 ETH, or about $96,000 per node, whereas Proof of Work requires a mining setup of less than $10,000. This means Ethereum is far less democratized. There was also a report in May of 2024 that one whale staked about $500 million to the network. There is also some centralization by the very fact Lido has such a large pool of validators at 33%.

Solana receives less criticism as it offers Proof of History (PoH) which offers a high throughput of 65,000 transactions per second on GPUs although other blockchain networks have a higher time to finality. Solana accomplishes this without second layers or side chains by using the Proof of History (PoH) consensus mechanism. PoH creates a historical record that proves an event has occurred at a specific moment in time. Rather than validators agreeing on a time, Solana validators maintain their own clock by encoding the time with a cryptographic hash function (SHA-256). This circumvents the need to wait for confirmation, thereby resulting in a higher throughput.

The purpose of this discussion is to say that many Layer 1s need Layer 2s or middleware to solve for the third tenet of the blockchain. For Ethereum, this is scalability (yet some critics also point toward a lack of decentralization compared to what other Layer 1s offer). What is unique about Layer 2s is the addressable market may be higher than a Layer 1, as the space is heavily fragmented with many Layer 1s competing. The very best middleware and Layer 2s can solve a critical pillar for these competing Layer 1s with little to no fragmentation.

Quick Primer on Smart Contracts

Decentralized systems require contracts if a system is to be created where all data, all messages, all token transfers, and all users are validated. Machines need contracts to offer ultimate security and decentralization as contracts ultimately allow for a revocation if a request is found to be fraudulent. In this case, the machine can shut the request down or otherwise deny the user/token transfer/data/message to move forward. The bilateral nature of contracts allows for a validation process to where nodes can determine if the action is trustworthy.

To put it simply, without contracts, trusted systems cannot truly exist as otherwise there is no penalty. This is why Web2, which is not based on smart contracts, is rife with bad actors. One centralized system can create thousands of bots, or a centralized tech company can push its agenda to the top of a newsfeed. There is no contract, and therefore, there is no revocation for unethical behavior.

Chainlink solves the critical issue of supplying trusted contractual data for smart contracts. Oracles are trusted third-parties that retrieve off-chain information and push that information to the blockchain at predetermined times. Technically, oracles introduce a potential point of failure, and this is why Chainlink’s ultimate goal is to maintain a good reputation. With a good reputation, the network effect will take care of itself since blockchain developers will be attracted to whichever oracle network is the most reliable.

Solving for Decentralization with Oracles:

Decentralization for the blockchain is achieved by having backend code on a decentralized network instead of a centralized server. Developers use a blockchain like Ethereum for data storage and use smart contracts for the app logic. The Ethereum blockchain is run off of thousands of nodes. These nodes are constantly computing the transactions within the blockchain from around the world, making it nearly impossible to hack as well as regulate.

Decentralized applications (Dapps) rely on smart contracts. Dapps deployed on the Ethereum network are controlled by logic written into the smart contract and cannot be altered by the developer. Smart contracts function like APIs (this was also discussed in the Chainlink webinar). This allows applications to build on one another similar to the way applications use APIs today; except blockchain applications build on smart contracts.

Chainlink was primarily built for off chain data for non-currency smart contracts. The principal is the same where there is a set of rules which self-execute – the more common analogy is that it operates like a vending machine to where there is no middleman. When you use a vending machine, you’re inputting a payment and terms (like pushing the buttons #D5) and the output is a bottle of water; the exchange did not require a gas station clerk. Smart contracts are similar in that terms are defined, and when those terms are met, there is an output.

The front-end application can be written in any language with calls made to the backend. The main qualities are that the applications are decentralized, can perform any action given the required resources (whereas Bitcoin is not Turing complete – more on this below) and are executed in a virtual environment such as the Ethereum Virtual Machine. The virtual machine acts as a buffer to where if the application is faulty, it does not affect the blockchain network.

Sounds great, but There’s One Problem …

The concept of decentralization sounds great in theory, yet the problem remains that even if a network is decentralized and secure, the data the applications use may not be decentralized or secure. If banks use the blockchain to drive down costs associated with money transfers, where will the banks get secure foreign exchange data that can’t be tampered with?

Where the Bitcoin protocol differs from networks like Ethereum, Solana, Cardano, Polkadot, Avalanche and others, is that the Bitcoin protocol has only one purpose – which is to transfer and store Bitcoin. Bitcoin is not intended to function like an operating system, and developers cannot develop dencentralized apps (Dapps) for the Bitcoin protocol. Due to having a single purpose, the Bitcoin protocol does not introduce or rely on off-chain data.

As we stated in our 2019 writeup, it’s the world’s most secure network, and in fact, is more secure than 10,000 banks combined. This inherent quality is a primary thesis to the investability and moat of Bitcoin:

“Bitcoin is based on the most secure network in the world, and this solves a very real need for the financial system – which cannot be automated without a decentralized blockchain solution. Technically speaking, bitcoin is also the world’s most secure financial network. The transfers eliminate 3% in processing fees and hedges against inflation. This can, and should be, worth as much as a search engine, enterprise software, a social media network, warehouse fulfillment (AMZN) or iPhone hardware.

Apple, Google, Microsoft and Amazon reached market caps of $1 trillion because their products scale to global populations and are required on a daily basis. Bitcoin not only scales to the global population but it also protects their livelihood – a necessity rather than a convenience. In fact, we see populations who are not necessarily tech savvy most enthusiastic about bitcoin, and this is a strong signal that it will scale beyond the reach of $1 trillion market cap.”

Bitcoin is very different from other blockchain networks because it’s not Turing complete, which means the Bitcoin protocol is not intended to perform computations or to solve complex problems. By limiting its functionality, Bitcoin’s protocol is highly secure. Bitcoin Script verifies transactions but is not programmable for general-purpose computations. Rather than offer scripting capabilities, Bitcoin Script offers a restricted set of operations that apply only to transactions.

The Turing incomplete design offers the ultimate security rather than seeking the functionality of a general-purpose platform. There are about 50,000 distributed Bitcoin nodes, as well, that help to verify and record transactions with a decentralized architecture.

Bitcoin’s goal is very different from Web3 technologies, and this distinction is important. Ethereum’s Solidity programming language is designed to facilitate smart contracts on the Ethereum blockchain. It’s considered “Turing Complete” as it has key features such as reading and writing to memory, branching to move the machine forward and looping to restart execution.

Yet, decentralized applications will require some data, so how can a network claim to be decentralized if the data itself is not originating from a decentralized network?Yet, decentralized applications will require some data, so how can a network claim to be decentralized if the data itself is not originating from a decentralized network?

Swift is a system used by banks and institutions for money and security transfers. Currently, 11,000 banks use SWIFT across 200 countries, and it’s the largest and most streamlined payment system for international transfers. SWIFT facilities secure and efficient communication between institutions with 45 million messages sent in 2022. The reason Swift is looking to innovate with blockchain technologies is to avoid becoming disrupted. The company charges between $15 to $50 on average for a transfer and it can take hours or days for the transfer to be completed. The security and decentralization of a blockchain network is attractive to Swift, yet there will be issues if the pricing data used by the blockchain for settlement is not accurate.

In this example, Chainlink provides a pricing oracle for Swift’s smart contracts and for many other financial applications that need accurate pricing information for settlements. The pricing information needed for foreign exchanges, cryptocurrency swaps, stock prices and other assets is sourced from decentralized data.

Oracles assist in helping blockchains use off-chain data.

Chainlink has built a decentralized oracle network (DON) that aggregates independent node operators, reliable data sources and also oracle networks for accurate, decentralized data. For example, Chainlink will determine crypto pricing by combining many sources from price data aggregators (like Coinbase and dozens of others like Coinbase), with node operators (there are hundreds), and with oracle networks like Ethereum or Solana. By aggregating many data sources and many types of data sources, the pricing feed eliminates a single source of failure.

Solving for Scale: Chainlink Securely Enables Rollups

Blockchain seeks to disrupt nearly every industry, and yet, it’s unfathomable for some investors to envision this outcome when it’s costly and slow to transfer tokens and data. Ethereum’s primary issue is scale as the Layer 1 has seen exorbitant gas fees. There was an outlandish case where someone bought a NFT for 1 ETH, or about $2600, and was charged $70,000 due to high gas fees. Yuga Labs has seen gas fees of up to $176 million for $285 million in sales prior to the merge to Proof of Stake. Gas fees are a result of Ethereum having low transactions per second (TPS) in the 20-40 transactions range. Once the limit is reached, the remaining transactions compete, resulting in higher fees.

Proof of Stake (PoS) relies on a few key technologies to drive down energy consumption, such as staking with validators for decentralization purposes, shards that break tasks into a subset, a dispersed network of nodes to increase efficient processing using the Beacon chain, and Rollups where hundreds of transfers can be rolled into a single transaction. With Rollups, the smart contract verifies all of the transfers in the Rollups. The goal is to reduce computing and storage resources by reducing the amount of data held in a transaction.

Ethereum offers ZK Rollups, or Zero-Knowledge, which we’ve covered here. However, rollups are a dedicated instance, and therefore can be customized and are interchangeable with other networks. A company called Cosmos released Rollkit and other developer tools that increase time to market for developers while seeking to lower the cost of zero-knowledge proofs.

Therefore, the development and functionality are separate from the Layer 1. This is especially attractive to enterprises that do not need a specific Layer 1 (like Ethereum) to reach scale since they already have a customer base. We touched on Coinbase using Rollups for the Base Layer 2 stating it offers “1 cent, 1 second transactions.” SAP has recently minted USDC on a Layer 2 solution with a statement it’s “getting close to that one cent cost basis.”

The Optimistic Collective developed a network and stack that Coinbase’s Base uses for its Rollups. This means that CB’s Base settles on Ethereum’s network for security purposes, yet uses the OP stack to execute Rollups outside of Ethereum. This is ideal as Coinbase can switch blockchain Layer 1s at any time. The amount of development that has occurred on Layer 2s since Rollups and the merge to PoS on Ethereum is staggering. For example, according to Chainlink, there are “approximately 48X the amount of on-chain calls from dApps to Chainlink Data Feeds over the past 12 months compared to Ethereum’s baselayer.”

Offchain Labs created a Layer 2 called Arbitrum that directly competes with CB’s Base and is currently ahead of Base on a few key metrics, although Base’s growth has been more rapid and will likely surpass Arbitrum. According to the writeup and data from Token Terminal, Arbitrum’s fees are much lower than Base, which is a distinct advantage in utility, yet Base is leading on gross profits at $30 million estimated compared to Arbitrum’s $9.5 million.

With Rollups, Chainlink’s purpose has expanded to where Chainlink nodes are used to do computations and verifications before posting a bundled transaction. This means an app built on a Layer 2 like Arbitrum or Base only has to use the main chain when necessary.

Layer 2 chains and rollups have a separate ledger, and there’s a chance a malicious actor forks a Layer 2 to create a parallel chain to mint new assets or burn assets on the Layer 1. Decentralized oracle networks like Chainlink prevent this from happening by offering Oracle Network Governance, or a trusted source, for routing upgrades to L2/Rollup chains, importing the ledger state, resolving disputes, and helps facilitate L2s that are forked.

Rollups provide an easier path to development as enterprises can avoid having to build a complex Layer 1 chain, while also providing much faster and cheaper transactions. This is ultimately a catalyst for Chainlink to become the oracle network for enterprise dApps. For example, Sony Group’s venture arm recently announced the company is using Chainlink to launch a developer platform that delivers consumer blockchain applications. Similar to Base, Sony will also use Optimism for its scaling layer, while using Chainlink for Data Feeds.

There are 200 protocols that operate on Base (and growing). Chainlink Automation allows developers to offload tasks that require high computational power to the Chainlink Network for a reduction in fees by 90%. Chainlink’s Cross-chain interoperability allows for currency and data transfers between blockchain networks, which can help with Base’s goal of having low cost, yet fast transfers.

Section 2: How Chainlink is Becoming a Blockchain Frontrunner

Here are some of the technologies that Chainlink is using today to quietly become a frontrunner in the crowded and complex Blockchain space:

Chainlink is an abstraction layer, which means it’s agnostic and works across all major chains, such as Polkadot, Avalanche, Binance Smart Chain, Polygon, Optimism, Arbitrum, and more, which eliminates vendor lock-in. This is incredibly important given how nascent the Blockchain ecosystem is, and how the rate of failure is already high and will remain high as blockchain technologies continue to mature.

Chainlink is also future-proofed, which refers to being compatible and scaling even as the blockchain is continually developed. Amazon Web Services (AWS) offers a Quickstart for Chainlink to where DevOps teams can quickly launch a Chainlink oracle node on AWS to sell real-world data. The framework is future-proofed because Chainlink continually updates the integration for data providers to sign their data and broadcast it to the blockchain. For example, NOAA hosts weather data on Google Cloud, which can then be accessed on Ethereum’s blockchain, which in turn powers crop insurance agreements.

Trust-Minimized Oracle Computation: As enterprise apps grow in complexity, Chainlink not only offers trusted off-chain data through decentralized oracle networks (DONs) but the company also offers developers a path in performing off-chain computations. These off-chain computations are gasless and as fast as native hardware. For example, if a developer wants to make a conclusion based on historical weather data, there would be multiple steps in taking the API data and performing a computation. These computations can be done on behalf of smart contracts in a trust-minimized manner. Chainlink Functions is a serverless developer platform that offers compute runtime to test, simulate and run offchain logic for Web3 apps, similar to how AWS Lambda offers serverless solutions.

Chainlink Automation: Automation uses oracle computation to run predefined conditions and to trigger smart contracts, for example, when stock or token limit orders are hit on a decentralized exchange. The decentralized network of nodes performs the off-chain computation of the contract’s logic that is then verified on-chain.

According to Chainlink, there are three key benefits to using the company’s smart contract automation. The first is that the company removes any centralized point of failure through decentralized oracle networks (DONs). Secondly, DevOps time is reduced by leveraging LINK’s Automation-compatible contract infrastructure. Third, developers enhance the security of their protocols by removing the need for centralized servers.

Secondly, rather than investing time and resources in creating scripts for on-chain monitoring and triggering smart contract execution, developers can plug into Chainlink Automation’s optimized infrastructure by simply creating an Automation-compatible contract and registering it. This saves time, reduces the DevOps workload, and allows developers to focus on writing more great code. 

Lastly, by using Chainlink Automation, developers can enhance the security of their protocol. Developers no longer have to risk exposing their own private key when initiating transactions from centralized servers—the nodes on the Chainlink Automation Network will sign on-chain transactions.

Cross-Chain Interoperability Protocol (CCIP): Blockchains are fragmented and there is too much friction in payment transfers and information exchanged between applications. Web3 must function seamlessly like Web2 to where the infrastructure (AWS, Azure, GCP), the protocols (TCP/IP, SMTP) the operating systems (Windows, Linux, MacOS, Android, iOS), the applications and the software are seamlessly working together without friction. In the majority of these cases, the user is unaware of the systems that support the user experience.

Chainlink’s CCIP sets out to solve this by providing a bridge between blockchains and DeFi applications. The protocol was launched in July of 2023 to solve the pain point of seamlessly transferring data and currencies across various blockchain networks. At launch, it was integrated with Ethereum, Avalanche, Polygon and Optimism. This allows users to use any decentralized application (dApp) on these blockchains for liquidity purposes and connectivity.

Chainlink is uniquely positioned to solve the problem of bridging blockchains and popular applications because it has built a secure oracle network. CCIP extends the idea of an oracle network, which was originally designed to on-load off-chain data, to also offer decentralized oracle computation for performance histories and to monitor for malicious activity. Off-Chain Reporting (OCR) is used to aggregate a report from many validators, which reduces congestion.

Please note: The deep dive 2023 Chainlink Update: Interoperability for Blockchains, Bullish SWIFT Partnership goes into more detail around the importance of CCIP.2023 Chainlink Update: Interoperability for Blockchains, Bullish SWIFT Partnership goes into more detail around the importance of CCIP.

Price Feeds: Chainlink Price Feeds provide smart contracts with access to financial market data. Decentralized Finance (DeFi) is a $100 billion+ market and whichever oracle network is deemed most reliable will carve a deep moat for itself. Price Feeds are used for on-chain currency transfers where a reliable token or currency amount is needed. Chainlink’s data feeds and price feeds currently secure tens of billions in value across DeFi.

Chainlink SCALE: Chainlink’s SCALE program was launched to help newer blockchains to cover the cost of decentralized oracle networks initially before decentralized apps (dApps) can cover these costs. This helps to reduce the operating costs of oracles while allowing Layer 2s to access higher volumes of data and more advanced applications. This helps to incubate apps where startup costs would otherwise be prohibitively high.

Chainlink Verifiable Random Function (VRF): Blockchain engineers use random number generation for cryptographic keys to prevent tampering. Private and public cryptographic keys are used today to encrypt and decrypt data, and are used across domain name systems (DNS), application security and zero-knowledge trust. There is a similar need for private and public cryptographic keys for blockchain applications. Chainlink VRF is the most widely adopted random number generator (RNG) in Web3, and has generated over 21 million request transactions for thousands of smart contracts.

Chainlink’s Ecosystem:

Chainlink launched on the Ethereum mainnet on May 30th, 2019. We covered the altcoin for the first time in August of 2019. Prior to launch, Chainlink signed 30 partnerships. By 2020, when we covered Chainlink a second time, total value secured (TVS) had grown from $254 million to $6.3 billion, or 23,000% growth in TVS. At crypto’s peak in Nov of 2021, Chainlink had $75 billion in total value secured across up to 1500 protocols and hundreds of DeFi applications. This has declined to $25.5 billion today and $31 billion in the past 30 days, yet the TVS from 2021 is important as it shows Chainlink is capable of securing $75 billion without a hack. As the CEO stated at SmartCon:

“Right now, the Chainlink Network has provided the most cryptographic truth in history.”

According to CoinDesk, the transaction value of data points that Chainlink has delivered on-chain stands at $15 trillion, up from $9 trillion at the start of the year. As stated, there’s been no hacks or otherwise any tokens lost through Chainlink’s secure network.

The predominant use of Chainlink is for market data for DeFi apps. DeFi apps have grown in total value locked (TVL) from $700 million in December of 2019 to more than $90 billion today. With the launch of CCIP, Chainlink can use its strong track record in securing DeFi blockchain smart contracts to expand to cross-chain smart contracts.

Update on SWIFT Partnership:

Perhaps where Chainlink is most promising in the near-term is the SWIFT Partnership as Swift facilitates $50 billion in transactions every year.

SWIFT stands for the Society for Worldwide Interbank Financial Telecommunications (SWIFT) and is the system used by banks and institutions for money and security transfers. Currently, 11,000 banks use SWIFT across 200 countries, and it’s the largest and most streamlined payment system for international transfers. SWIFT facilities secure and efficient communication between institutions with 45 million messages sent per day in 2022.

When we last covered the SWIFT partnership, a test had been completed using Chainlink’s CCIP to facilitate transactions with tokenized assets on public and private blockchains using back-end systems. This will allow financial institutions to integrate blockchain technology into the existing infrastructure. You can read more about this on Swift’s website here.

This month, it was announced the testing phase has been complete and the Swift partnership is now in the pre-production phase.

Tokenomics:

One area where critics find fault with Chainlink is the tokenomics. The monthly growth in Chainlink’s circulating supply is at 1.4% per month on average. Over the course of a year, this can dilute token holders 10 to 15%, on average.

Chainlink’s fully diluted market cap tracks 2.1X higher than its current market cap. There is a circulating supply of 568M tokens yet a max supply of 1 billion tokens.

Of the 1 billion tokens, at the initial coin offering, a little more than one-third was to go to node operators, a little more than one-third was sold in the public sale, and a little less than one-third went to the company to be held in reserves. The 350 million held at the company can be released anytime, which dilutes token holders.

When more of Chainlink’s supply is in circulation, there is likely to be stronger price action. About 63% is in circulation now, and so look for Chainlink’s price to be less volatile in 2 years if we assume the 1.4% per month rate in circulating supply continues.

Conclusion:

The blockchain is needed to decentralize information to where the global population is not dependent on Big Tech companies for data and inputs. I’ve boldly stated that if people are not concerned by this today, they are not paying attention. The internet is broken as tech companies, media companies and other corporations are full of self-interest and have proven they can’t be trusted. The internet has gatekeepers, and there is ample reason to remove these structures.

Looking beyond information from search and social media, of which there is a great dependency, there is also a dire need for the decentralization of loans, interest-bearing accounts and credit. Imagine if you could make the 20% APR that Chase makes off loaning for credit cards, yet to do so by using the blockchain for ensuring the borrower has a high credit score (or collateral) and through a smart contract to where the collateral is retrieved if the borrower does not pay you. Instead, we are offered 4% interest rates on money market accounts and CDs, and this is rare — for the past decade, it’s been as low as 0%. Finance is broken today, as it’s highly centralized and skewed in favor of a few.

There is also the upcoming need over the next decade to improve automation as machine-to-machine communication is not truly possible unless it encompasses the three tenets of the blockchain (security, decentralization, scalability). As the AI era evolves, this will become a driving force for blockchain use cases.

A Note on Crypto …

Last year, I was on a Real Vision interview where they asked me what my investment framework for crypto is and I said: “Technicals, technicals, technicals.” If you want to buy Apple or Microsoft without using technicals, and hold over a long period, that will probably work out just fine. But to participate in these extraordinary companies at an early stage, it’s of ample importance to carefully consider technicals.

We lead with technicals on crypto given it’s early-stage tech. This is different than stocks, where fundamentals lead. The good news is that crypto is sentiment driven, and so it respects price and technicals quite well when managing these positions. With that said, crypto is high risk, high reward and requires an active stance – it is not for those who are new to investing or coming to grips with the ups and downs of the market. We believe there will be extraordinary returns on Chainlink, yet there will also be extraordinary drawdowns.

Our history with Chainlink is quite good – we bought at $1.50 and trimmed in the $25 to $50 range, and then began buying again much lower in the $7 to $11 range. We plan to actively manage Chainlink moving forward. Please join Knox’s weekly webinars on Thursdays at 4:30 p.m. Eastern to hear more on how we plan to manage Chainlink, with a special emphasis this week on crypto.

Recommended Reading:

• Cloudflare Q3 24: Soft Q4 Guide as Company Transitions on Billing Terms
• Crypto and AI Opportunity: Real Vision Video Interview
• Real Vision Video Interview: Will Nvidia Continue to Dominate AI?
• 2023 Chainlink Update: Interoperability for Blockchains, Bullish SWIFT Partnership

Cloudflare beat on the top and bottom lines in Q3, although net retention rate softened again. The company missed on revenue expectations for Q4 with management guiding for a midpoint of $451.5 million, for growth of 24.6% compared to analyst expectations of $455.08 for growth of 25.8%. Management spoke at length about being at an inflection point for a sales organization shift, yet this felt like a speculative discussion as evidence of an inflection was not visible in the Q4 guide.

Key metrics were mixed and revenue growth is getting further away from the coveted 30% mark with a 4-point deceleration in Q4. Yet RPO and customer growth remains persistently strong.

Revenue

Cloudflare reported revenue of $430.1 million in Q3, increasing 28.2% YoY, slowing from the 30.0% YoY increase reported in Q2. Revenue growth had been expected to be steady at ~26% YoY for the next few quarters, which is quite strong for best-of-breed cloud as many >40% revenue growth cloud companies have dipped <20%.

However, Cloudflare guided to just 24.6% YoY growth in Q4, seeing revenue between $451 million and $452 million, shy of the estimate for $455.8 million. This implies a sequential deceleration of ~3.6 points.

For FY24, management slightly raised its guidance, though it seems as if the raise stemmed solely from Q3’s outperformance relative to its initial guide. Cloudflare now sees FY24 revenue of $1.661 to $1.662 billion, a raise of just $3.5 million at the midpoint from its prior view for $1.657 to $1.659 billion. Q3’s $430.1 million sum beat management’s internal guidance by ~$6.5 million, implying that it is the primary factor behind the raised guide given Q4 was a miss.

Margins

Cloudflare reported strong adjusted operating margin of 14.8%, which led to raising its adjusted operating margin guide for FY24 to 13.3%, up from last quarter’s FY24 guide of 11.9%. GAAP operating and net margins continue to make slow progress towards positive territory due to high stock-based compensation of 20.5%.

• GAAP gross margin was 77.7%, down slightly from 77.8% in Q2 but up 1 point from 76.7% in the year ago quarter. Adjusted gross margin was 78.8%, down from 79.0% in Q2 but up slightly from 78.7% in the year ago quarter.
• GAAP operating margin was (7.2%), improving from (8.7%) in Q2 and (11.7%) in the year ago quarter. Adjusted operating margin was 14.8% in Q3, well above the 11.9% guided and improving from 14.2% in Q2 and 12.7% in the year ago quarter.
• For Q4, management guided adjusted operating margin of 11.9%.
• For FY24, management guided for adjusted operating margin of 13.3%, a strong increase from 9.8% at the beginning of FY24 and its previous guide for 11.9% given in Q2. The increased guide is driven by Q2 and Q3’s outperformance, with both quarters seeing adjusted operating margin >14%.
• GAAP net margin was (3.6%), improving slightly from (3.8%) in Q2 and (7%) in the year ago quarter. Adjusted net margin was 16.9%, falling from 17.3% in Q2 but up slightly from 16.5% in the year ago quarter.

EPS

For fiscal year 2024, Cloudflare is expected to report EPS growth of 51% YoY to $0.74. Cloudflare beat slightly on EPS in Q3, while guiding for flat QoQ growth for adjusted EPS for Q4.

Cloudflare is close to GAAP profitability yet it’s not clear given stock-based compensation when the company will tip over the edge. The stated in the investor’s presentation the long-term goal is a 20% adjusted operating margin, which implies GAAP profitability would be narrow, yet achievable. Notably, there is no firm date being provided.

• GAAP EPS was ($0.04), flat QoQ and up from ($0.07) last year.
• Adjusted EPS was $0.20, beating estimates and Cloudflare’s guide for $0.18. This was also flat QoQ and up 25% YoY.
• For Q4, management guided for adjusted EPS of $0.18. For FY24, based on the slight beat in Q3, management raised its adjusted EPS guide to $0.74, up from its previous view for $0.70 to $0.71. This represents YoY growth of 51%.

Cash and Balance Sheet

Cash flow generation remained steady in the quarter as cash flow margins expanded sequentially.

• Operating cash flow was $104.7 million in Q3, for a 20% margin. This improved from a 19% margin in both Q1 and Q2.
• Free cash flow was $43.7 million, for an 11% margin, an improvement from 10% in Q2 and 9% in Q1. The company stated its goal is a 25% FCF margin.
• Cash and available-for-sale securities totaled $1.824 billion while convertible notes totaled $1.286 billion.

Cloudflare’s free cash flow is especially impressive given the company has to build a bigger network and invest in GPUs for edge AI.

“Network CapEx represented 10% of revenue in the third quarter. During the quarter, we saw a notable shift in customer conversations and buying behavior from AI training to AI inference, including our first multimillion dollar workers AI contract. 

This gives us confidence to continue to increase our investment in higher-end GPUs as well as the breadth of our GPU rollout as we provision greater capacity to support demand in 2025. As. A result, we continue to expect network CapEx to increase again in the fourth quarter to reach 10% to 12% of revenue for the full year 2024.”

Key Metrics:

RPO

RPO came in at $1.53 billion for growth of 39% year-over-year and up 6% sequentially. This is higher than the year ago quarter when RPO was at 30% YoY and up 5% sequentially.

Compared to last quarter, RPO reported growth of 37% year-over-year and 6% QoQ, for 69% of total RPO, which means this was technically a stronger quarter on this key metric.

DBNRR

Cloudflare’s DBNRR dropped again in Q3, falling to 110%, compared to 112% in Q2 and 115% in Q1. Throughout 2023, DBNRR hovered around 115% and higher, so the decline over the past two quarters is important to watch. This quarter, management reiterated the decline is being driven by slower net expansion in the larger customer cohorts as they move to “pool of funds” contracts, which shifts billings from annual contracts to pool of funds accounts that are on a monthly basis for three or more years (for larger customers).

Per management again this quarter:

“Our dollar-based net retention was 110%, down 2 percentage points quarter-over-quarter. While customer churn remains consistently low, our shift to more pool of funds deals with our largest customers, which represented nearly 10% of new ACV booked in the quarter, up from 1% a year ago has put downward pressure on dollar-based net retention and change the shape of revenue recognition in the short term. 

Over the long term, however, we believe pool of fund deals are very positive for the business as they represent our largest customers making a broad commitment to Cloudflare's overall platform.”

Customers

Paying customers increased 22% YoY to 221,540, accelerating from 21% in Q2 and 17% in Q1. Management stated on the call they count 35% of the Fortune 500 as customers.

For customers with >$100K ARR, growth decelerated 2 points sequentially to 28% YoY to 3,265 customers. This cohort accounted for 67% of revenue in Q3, flat with Q2 but up from 65% last year. Management pointed out that the 28% resulted in a record addition 219 large customers, referring to the growth remaining high even on a larger base of customers.

Billings

Despite other metrics such as >$100K ARR customers and DBNRR decelerating, Billings accelerated sequentially in Q3. Billings increased 24% YoY to $447.3 million, a 1-point acceleration from 23% growth last quarter.

Earnings Call:

Net Retention Rate; New Sales Org

The tone on the earnings call was more of a “wait and see” tone about the new sales organization reaching an inflection point. On one hand, we are seeing strong total customer growth and billings inflect. Yet, on the other hand, Cloudflare is asking investors to place their trust in the company on the “pool of funds” deals that are causing a rapidly decelerating net retention rate.

What investors should keep in mind is that management is foreshadowing the mixed key metrics could last for a few quarters as current customers shift to the new billing terms: “As we mentioned last quarter, we expect new customers to contribute a higher percentage of our overall year-over-year revenue growth for the next several quarters.”

There was also a note that larger deals were pushed back, yet the miss in Q4 doesn’t inspire confidence they were pushed back only by a quarter: “However, some larger deals slipped out of the quarter in the U.S. in particular, during what was a transitional period under new sales leadership in that region. These deals are still active in our pipeline with many having already closed this quarter.”

The company discussed being at an inflection point, yet as an investor, I prefer more evidence of this with revenue growth inflecting, as well (which did not happen): “All the changes in our sales force may have impacted the short-term cadence of some larger deal cycles. What stood out to me is that the third quarter felt like we hit the inflection point in the rebuild of our go-to-market team.”

However, when pressed, management remained confident it was truly an inflection point. My best guess is the inflection will catch up to revenue perhaps by Q2 given the note it’s a few quarters out.

Timothy Horan   Oppenheimer & Co. Inc.

There's a lot of moving parts, obviously, with the sales productivity and limited capacity there in the pooling. Can you maybe talk about the timing of when revenue growth can accelerate again. Do you think the fourth quarter around 25% guide, is that the bottom? Or do you think it's still a few more quarters out? And related to this, can you update us on what you think the timing of for the $5 billion revenue target that you have? And I had a quick product follow-up.

Thomas Seifert   CFO

Yes. Thank you for the question. As we said before, for us, in our subscription business model, revenue is very much a lagging metric. Sales capacities is a product of the amount of headcount we have and the productivity progress they are making, this translates into pipeline and sales prospects, turns into ACV and then ACV is recognized ratably over the lifetime of the contract as revenue. So it's very much a lagging indicator. And as we said before, models like this, there are slow on their way down, but they're also slowing their way up. 

But the important part is, as you heard in Matthew's prepared remarks that we think we have reached this key inflection point with net sales capacity now, which is the leading growth indicator have been patent. So from there on, we expect sales activity translating the ACV moving forward and going up. And you see this already in those parts of the world where this conversion and transition has happened successfully. Revenue was up 38% already in APAC, and it was up 1% in Europe, which is our highest productivity region has been over the last several quarters. So we think we have bottomed out from a net sales capacity perspective, and move forward from there.”

Capex Spending and Edge AI

Cloudflare has offered visibility in capex spending plans as it builds out GPU-powered edge servers for AI inferencing purposes. This involves large orders of GPUs, to which Cloudflare made it quite clear they are able to work with any hardware on the market and at a high utilization rate.

Given Cloudflare may one day become a leading AI stock, discussions around the AI market are provided for future reference as we along, especially given there is no official AI revenue number being reported by Cloudflare today.

Per the company, there has been progress: “During the quarter, we saw a notable shift in customer conversations and buying behavior from AI training to AI inference, including our first multimillion dollar workers AI contract.”

Notably, there was not an update to the Workers key metric of 2.4 million developers this quarter.

The discussions around one of Cloudflare’s key value propositions was the following:

“And in order to support that, we have made the investments to increase not only just the number, but also the power of the GPUs that we're deploying around the world. What I think is unique about Cloudflare is 2 things. One, we are actually able to deliver inference incredibly close to where anyone is on earth because we've deployed the inference capabilities across at this point, nearly all of our network. 

But in addition to that, we've actually done the work to get higher utilization out of those same GPU resources where what we see when we survey customers that are trying to manage this themselves, through hyperscale public cloud is that they're getting utilization rates that are sort of in the 5% to 10% range of the resources that they're buying.

We're able to deliver much higher utilization. And in the process of that, that means that we can actually pass on the effective savings to our customers. So they not only save in not having to maintain their own team to manage these virtual machines and containers, but they also save because we can just do more with the same GPU resources that are being deployed.”

Conclusion:

Cloudflare is among the highest in terms of valuations for best-of-breed stocks at a 19.7 forward PS. Although this was a decent report with no major red flags, this valuation is not where I’d be buying the stock given the report was not a knockout. The question being considered is if we should close the position since odds are quite high we will get the stock lower.

Recommended Reading:

• Cloudflare Q3 Earnings Preview: All eyes on the guide
• Real Vision Video Interview: Will Nvidia Continue to Dominate AI?
• AMD Q3 2024: GPU Revenue at 22%, and AI PCs have a Leader
• Micron Q4: Data Center Drives Beat, Profitability Soars

Cloudflare will release its Q3 results on November 07th. Analysts expect revenue to grow 26.4% YoY to $424.12 million and adjusted EPS to grow 14.2% YoY to $0.18.

During Q2 results, management increased the FY revenue guide to $1.658 billion at the mid-point from $1.65 billion, representing YoY growth of 27.9%. The FY2024 adjusted operating margin guide was raised to 11.9% from 9.8%, up from 9.4% in 2023.

There was a survey published by Wells Fargo that stated Cloudflare is taking more market share from cybersecurity vendors. The downside is the report also stated Q3 was the weakest environment they’ve seen in three years. We need an earnings report to confirm the results of the survey, yet it’s encouraging to see NET is stronger than its peers in a spending environment that is largely outside of the company’s control.

Regarding AI, Cloudflare is uniquely positioned to capture inference at the edge. The company has grown developers on the Workers platform from 2 million to 2.4 million, for 20% QoQ growth. Companies with decent sized developer moats (Nvidia, Apple, etc) have a developer following in the 3 to 4 million range. Given the majority of the AI market is focused on training large language models at the moment, inference needs time, but there are early signs of which companies will succeed as AI matures.

Revenue

The company’s revenue is expected to grow steadily at around 26% in the next few quarters. The growth is quite strong for best-of-breed cloud as many >40% revenue growth cloud companies have dipped <20% in recent years in what has been a seismic shift in cloud. Our firm was early to report on this shift for our members. Yet, Cloudflare is one of the very few that have sustained strong growth levels.

• Q2 revenue grew by 30% YoY to $401 million. Management guide for Q3 is $423 million to $424 million, representing a 26.2% year-over-year growth at the midpoint.
• Analysts expect Q3 revenue to grow 26.4% YoY to $424.12 million and 25.8% YoY to $455.80 million in Q4.

• Last quarter, management increased FY2024 revenue guide to $1.658 billion at the midpoint from $1.65 billion, representing 27.9% YoY growth.
• Analysts expect revenue to remain steady over the next three years at 28% growth, 27.1% growth and 27.8% growth YoY through 2027.

Margins

Cloudflare showed strong improvement in operating margin and net margin in Q2. It also significantly increased the full-year adjusted operating margin guide from 9.8% to 11.9%.

• Q2 gross margin was 77.8% compared to 75.6% in the same period last year. Adjusted gross margin was 79% compared to 77.7% in the same period last year.
• Operating margin was (-8.7%) compared to (-18.2%) in the same period last year. Adjusted operating margin also significantly improved to 14.2% from 6.6% in the same period last year. The operating expenses were reduced due to focus on higher productivity and better efficiency in the operations.
• Management’s adjusted operating income guide for next quarter is $50.5 million at the midpoint or 11.9% of revenue compared to $42.5 million or 12.7% of revenue in the same period last year.
• The difference between the GAAP and non-GAAP operating margin is due to stock-based compensation which was $86 million in Q2 or 21.4% of revenue. The high level of stock-based compensation reflects what the competitive cloud industry must do to retain talent.
• During Q2 results, management had increased the FY 2024 adjusted operating income guide from the range $160 million-$164 million or 9.8% of revenue at the mid-point to $196 million-$198 million or 11.9% of revenue at the midpoint. By doing the math, the Q4 adjusted operating income guide comes to $47.1 million or 10.4% of revenue.
• Q2 adjusted net income was $69.5 million or 17.3% of revenue compared to $33.7 million or 10.9% of revenue in the same period last year.

EPS

Analysts have increased adjusted EPS estimates after the company raised the FY 2024 guidance to $0.70 to $0.71 from the earlier guide of $0.60 to $0.61. Management Q3 adjusted EPS guide is $0.18.

• Analysts expect Q3 adjusted EPS to grow 14.2% YoY to $0.18 and 15.4% YoY to $0.17 for Q4.
• Analysts expect strong growth with 2024 adjusted EPS to grow 46.1% YoY, followed by 20.5% in 2025, and 28% in 2026.

Cash Flow and Balance Sheet

The company’s cash flows are improving. Management has reiterated they expect full year free cash flow in the range of $160 million to $164 million. This is a significant improvement from the $119.5 million free cash flow in 2023 and also suggests an acceleration in the second half with estimated free cash flow of $88.1 million.

• Q2 operating cash flow was $74.8 million or 19% of revenue compared to $64.45 million or 21% of revenue in the same period last year.
• Q2 free cash flow was $38.3 million or 10% of revenue compared to $19.97 million or 6% of revenue in the same period last year.
• Network capex was 6% of revenue in Q2. Management expects network capex to reach 10% to 12% of FY 2024 revenue in the 2H of the year as the company is rolling out GPUs in every location. We see setting these expectations while ultimately increasing cash flows as a positive.
• The company had cash and available-for-sale securities of $1.757 billion and debt of $1.285 billion compared to $1.716 billion and $1.284 billion in Q1.

Key Metrics

Remaining Performance Obligations (RPO) Accel’d QoQ

RPO increased 6% sequentially and 37% YoY to $1.42 billion although it was a deceleration from 40% growth in Q1. The current RPO was 69% of total RPO.

Billings

The company primarily focuses on RPO as a more comprehensive measure of its business. We track billings since they are reported for other cybersecurity stocks. Billings grew by 23% YoY and 9% QoQ to $421.7 million, a slight deceleration from 24% growth in Q1.

DBNRR

DBNRR was 112% in Q2, a 3-percentage point deceleration from 115% in Q1. The CEO stated the decline was driven by slower net expansion in the larger customer cohorts and an increase in “pool of funds” contracts. The pool of funds contracts is a transition in billing from annual contracts to pool of funds accounts that are on a monthly basis for three or more years. The pool of funds accounts are unique to the largest customers (for example, 4 of the top 10 customers are this account type) that use many products across the entire Cloudflare platform. These are considered larger platform deals that are paid on a monthly basis in a multi-year contract rather than an annual contract on one product. This is shifting how DBNRR and RPO are reported since revenue is recognized as the customer consumes the service.

Customers and Workers AI platform

Paying customers grew by 21% YoY to 210,166 and accelerated from 17% in Q1. For customers with an ARR of >$100K, Cloudflare reported 30% YoY growth to 3,046. This customer cohort contributed 67% of revenue, flat with Q1 yet up from 64% in the year ago quarter.

Cloudflare Worker Applications grew from 2M developers to 2.4M developers in four months, per the CEO’s opening remarks. The Workers AI Platform developer accounts grew 67% QoQ and inference requests grew 700% QoQ.

Other key point to watch

Customer Wins

Wells Fargo noted that the Q3 security reseller survey was the weakest in the last three years. However, Cloudflare’s survey results showed a strong uptick. “Cloudflare's (NET) results significantly up ticked to +31% net (-9% last qtr), as resellers noted strong overall demand trends for the full suite, including SASE, CDN, and cloud security.” This data point is not meant to make an earnings call, rather we will look for management to confirm the results of Q3.

Management mentioned in the last earnings call that the go-to-market initiatives are reaping rewards. The company also delivered another quarter of double-digit year-over-year increase in sales productivity during the second quarter, and saw an uptick in close rates, and an improved sales cycle.

The management also announced several large customer wins, and the notable one was the land & expand customer, who was a free customer: “A leading Australian technology company expanded their relationship with Cloudflare, signing a two-year $17.5 million contract, $7.2 million of which is expansion. They started with Cloudflare back in 2016 as a free customer and today use nearly all our products spanning use cases as diverse as remote application access, worker serverless development and bot management.”

Another prominent U.S. university signed a five-year $5.7 million contract. With Cloudflare, they were “able to replace multiple legacy vendors with a unified platform and cloud-first architecture.”

Valuation

Cloudflare is trading at a P/S ratio of 20.04 and a fwd P/S ratio of 18.06 and is trading below its five-year average P/S ratio of 32.67. Cloud has seen a re-rating, so the multi-year averages are not reliable in the current macro backdrop.

Conclusion

Given CrowdStrike has stumbled recently, Cloudflare has emerged as the leading best-of-breed cybersecurity stock. Although not GAAP profitable yet, the company saw a slim (3.8%) net margin in the last quarter, which puts GAAP profitability on the horizon. In the meantime, the company is improving its margins and cash flows, putting investor concerns to rest about capex, and is quietly sitting on one of the biggest developer ecosystems in the market with a fairly frictionless path to onboard the Workers AI platform once the inference market takes off.

We continue to watch Cloudflare’s report with anticipation as the company is firing on many cylinders. We are participating in the event Cloudflare beats, and we also have a trading plan in mind should we be able to get shares lower.

Royston Roche, Equity Analyst at the I/O Fund, contributed to this article.

Recommended Reading:

• Cloudflare Q2: Significant Margin Expansion, Customer Acceleration
• Real Vision Video Interview: Will Nvidia Continue to Dominate AI?
• AMD Q3 2024: GPU Revenue at 22%, and AI PCs have a Leader
• Optical Interconnects Overview: Strong Growth Expected Ahead