Ben’s Portfolio update end of March 2026
Returns and portfolio holdings:
Portfolio Notes 2022 -15.6%* *Jul-Dec, since I started posting my portfolio on Saul’s and fully adopting my version of Saul’s investing approach. 2023 77.8% 2024 31.7% 2025 24.9% 2026 YTD Month Jan -8.2% -8.2% Feb -15.9% -8.3% Mar -14.7% 1.3%
Timestamp: March 31, after market close
These are my current positions:
Mar 2026 Feb 2026 First buy* Cloudflare 25.1% 21.2% 11/2/2020 Nvidia 21.2% 21.8% 5/13/2020 Datadog 13.0% 12.5% 5/13/2020 Crowdstrike 10.1% 9.8% 5/13/2020 Snowflake 9.9% 11.2% 2/8/2021 Axon 7.4% 9.6% 4/2/2024 AppLovin 5.3% 5.9% 11/18/2025 Samsara 3.7% 3.4% 1/8/2024 Zscaler 2.2% 2.3% 3/4/2021 Astera Labs 2.1% 2.3% 11/18/2025
*held through today
Timestamp: March 31, after market close
Company comments
Nvidia:
Key insights:
- Nvidia keeps firing on all cylinders, comfortably exceeding my revenue expectation and for the first time since at least 2019, adding more absolute sequential revenue in a Q4 than in any preceding Q3.
- Delivered absolutely stunning new Q1 guide implying further YoY growth acceleration at an insane base, and updating my expectation to 84% YoY growth, from previously 63% YoY growth.
- 9 gigawatts of Blackwell infrastructure are already deployed and being consumed, while Hopper, and even six-year-old Ampere products, are completely sold out in the cloud. Blackwell is the “king of inference” due to its radically lower cost per token and perfectly positioned as computing demand for inference has grown 10000x over the last two years.
- Nvidia achieved a 35x throughput increase for high-value token generation. “Our Total Cost of Ownership is so good that even when the competitor’s chips are free, it’s not cheap enough”. The logic behind this argument is that in the world of $40b AI factories, the electricity and infrastructure costs of using inefficient free chips would cost a company more in the long run than when using Nvidia’s premium hardware.
- With the new Vera Rubin architecture, Nvidia projects a mind-bending 350x increase in token generation speed per gigawatt within just two years.
- Networking grew 3.5x YoY!
- Gross margins increased slightly to 75% showing the complex Blackwell supply chain and packaging yields have indeed stabilized quickly, and Nvidia is maintaining extreme pricing power.
- Finally, Huang doubled the $500b revenue expectation number which he gave in the previous quarter to now $1T (from CY2025 through CY2027 and for Blackwell and Rubin alone), which, adding in networking + Nvidia Groq integration + CPUs + Software, incl. enterprise AI + Sovereign AI + automotive + physical AI, leads me to a revenue growth expectation in between 85-90% YoY revenue growth for FY27 and 60-65% YoY revenue growth for FY28.
- If that becomes a reality and Nvidia stays at its current P/S ratio, the stock price would triple in the next two years.
Nvidia reported its Fiscal Q4 2026 earnings on 02/25/2026, delivering another phenomenal quarter that comfortably surpassed all my prior expectations. While I anticipated revenue of $67600M (18.6% QoQ, 71.9% YoY), the company achieved a record $68127M (19.5% QoQ, 73.2% YoY), delivering a 4.8% beat. This is quite a milestone and here is why: for the first time since at least 2019, Nvidia added more absolute sequential revenue in a Q4 than in any preceding Q3. And just when it seems the law of large numbers must catch up, management pushed their guidance higher. They guided Q1 FY27 to $78000M (14.5% QoQ, 77% YoY) which I now interpret as $81100M (19% QoQ, 84% YoY). It is kind of crazy to see a company accelerate its YoY growth from a $70b quarterly run-rate. And that, while entirely writing off the Chinese market. The China risk hasn’t just been de-risked as I noted in Q3; it has been completely neutralized and removed from the equation as a bearish talking point, while at the same time Sovereign AI brought in over $30 billion for the full fiscal year.
Looking at how the narrative has evolved from my Q3 assessment, the transition from Blackwell anticipation to full-scale realization is expanding; Management now expects growth to actually exceed what was included in that original $500B opportunity (more on that later). We knew Blackwell was off the charts in Q3, but in Q4 management confirmed that nearly 9 gigawatts of Blackwell infrastructure are already deployed and being consumed. What’s wild is that despite the Blackwell and Blackwell Ultra ramp, Hopper, and even six-year-old Ampere products,are completely sold out in the cloud.
We also have to talk about Networking: it is no longer just a supporting factor. Networking hit $11b in Q4 alone (up over 3.5x YoY), proving that the complexity of tying together 100K+ GPU clusters with Spectrum-X Ethernet and NVLink is a huge business in its own right.
In terms of profitability, the numbers tracked perfectly with my targets. I wanted to see GAAP gross margin around 74.8%, and they delivered 75.0%, alongside a 75.2% non-GAAP gross margin (versus my expectation of 75.0%). This confirms the bullish signal I noted in Q3: the complex Blackwell supply chain and packaging yields have indeed stabilized quickly, and Nvidia is maintaining extreme pricing power even as it scales new products. To put their operational efficiency into perspective, the company generated $96.6b in free cash flow for FY26 and returned $41.1b to shareholders via repurchases and dividends.
The outlook for FY27 and beyond builds on the Industrial AI thesis I discussed last quarter, but with a new twist: the enterprise adoption of AI agents. If FY26 was about laying down the foundational hardware for hyperscalers, the commentary suggests FY27/FY28 will see skyrocketing demand from enterprises deploying autonomous agents that require constant inference compute. Blackwell is already being positioned as the “king of inference” due to its radically lower cost per token. And even physical AI and robotics are actively contributing to the top line, adding over $6 billion in annual revenue. This proves that the pivot toward Industrial AI (factories, heavy industry, automotive) is already well underway, expanding the TAM.
In summary, Nvidia’s Q4 numbers pretty much killed any talk of a slowdown. Everyone from big tech and foreign governments to robotics startups is lining up for these chips. With supply constraints being navigated and great visibility into the future, the risk remains purely on execution, and Nvidia is executing without missing a beat.
And then came GTC. This is Nvidia’s yearly GPU technology conference which Jensen Huang kicked off with a keynote presentation that was not only very entertaining to watch (you should if you haven’t already), but also quite eye opening + I really like Jensen Huang’s positive, but realistic and grounded attitude towards AI; especially with all the doomsday talk going around.
As hinted to earlier, the narrative has shifted towards inference and the economics around monetizing inference. With the introduction of Open Claw and Nvidia’s enterprise ready version, called Nemo Claw, the industry is moving towards an “Agentic-as-a-Service” model (my guess is the corresponding acronym probably won’t stick, haha). The computing demand for inference has grown 10000x over the last two years, making data centers what Jensen calls "AI token factories”. And with the launch of the new Vera Rubin architecture, featuring the Vera CPU, 6th-generation NVLink 72, and Spectrum-X optical scale-out, Nvidia projects a mind-bending 350x increase in token generation speed per gigawatt data center within just two years. And, coming back to the topic of token monetization, Nvidia achieved a 35x throughput increase for high-value token generation. Just to put this into perspective Jensen dropped the hammer saying “Our Total Cost of Ownership is so good that even when the competitor’s chips are free, it’s not cheap enough”. The logic behind this argument is that in the world of $40b AI factories, the electricity and infrastructure costs of using inefficient free chips would cost a company more in the long run than when using Nvidia’s premium hardware.
Last quarter (Q3), the revenue opportunity for Blackwell and Rubin was pegged at $500 billion. After hinting at an even bigger number during the Q4 call, Jensen Huang used GTC to officially update that estimate. And I was quite shocked to hear him say that the $500b revenue opportunity from CY2025 through CY2027 became a $1T revenue figure, which will likely be exceeded. Keep in mind that this figure is for Blackwell and Rubin only, and my guess is that networking + Nvidia Groq integration + CPUs + Software, incl. enterprise AI + Sovereign AI + automotive + physical AI should increase this number further to something close to $1.25T. Using this figure to do some napkin math for the corresponding time frame of the last month of 4Q25 through two thirds of 4Q28, I am now penciling in between 85-90% YoY revenue growth for FY27 and 60-65% YoY revenue growth for FY28. If that becomes a reality and Nvidia stays at its current P/S ratio, the stock price would triple in the next two years. And with their supply chain already scaling up to multi-gigawatt output every month, the main risk that remains is, as I said earlier, execution.
Overview of how Nvidia performed versus my prior expectations:
- Revenue expectation: $67600M (18.6% QoQ, 71.9% YoY), implying a 4% beat; they beat a year ago’s Q1 guide by 8.5%, Q2 guide by 7.3%, Q3 guide by 7.9%, Q4 guide by 4.9%. Then last Q1 guide by only 2.5%, their Q2 guide by 3.9% and their Q3 guide by 5.6%.
→ $68127M (19.5% QoQ, 73.2% YoY), a 4.8% beat. Noteworthy: First time since at least 2019 that they added more sequential revenue in a Q4 than in the preceding Q3! - Q1 new revenue guide: $71666M (6% QoQ, 63% YoY) which I would interpret as $74370M (10% QoQ, 69% YoY), expecting slight YoY deceleration.
→ New Q1 guide: $78000M (14.5% QoQ, 77% YoY), which I now interpret as $81100M (19% QoQ, 84% YoY), implying a 4% beat. Kind of crazy to accelerate YoY growth (even without any guidance beat) from such a base … - I would like to see GAAP gross margin around 74.8%.
→ 75.0% GAAP gross margin. - I would like to see non-GAAP gross margin above 75.0%.
→ 75.2% non-GAAP gross margin. - Thoughts from previous quarter: Nvidia’s 3Q26 earnings recap.
Snowflake:
Key insights:
- first Q4 in 5 years, where sequential net new revenue didn’t drop seasonally 20-40% QoQ, but was essentially flat, demonstrating that underlying consumption trends are accelerating.
- Interpreting their Q1 and FY guides, product revenue growth will likely stay durable at 30% YoY for FY27.
- This is supported by cRPO growth accelerating to 36.4% YoY (up from 32% last quarter)
- Very strong customer growth rates, exceeding my prior expectation across all cohorts, including a record 56 customers spending over $10 million annually (a 56% YoY increase).
- “Snowflake Intelligence has scaled from a nascent offering to an essential capability for over 2500 accounts, almost doubling quarter-over-quarter”, showing the potential for a consumption explosion as consumption will shift away from human-centric work loads to AI agents.
- 61% FCF margin in Q4, operating income grew 50% YoY and free cash flow grew 85% YoY in Q4!! And FY26 operating income grew 111% YoY while FY26 free cash flow grew 27% YoY
- Stock-based compensation (SBC) is finally being turned down, dropping from 41% of revenue in FY25 to 34% in FY26, with a firm guide to drop to 27% in FY27.
- Their recent $600M acquisition of Observe brings Snowflake into the $50B IT operations and observability market.
Snowflake’s Fiscal Q4 2026 earnings report, released on 2/25/2026 successfully validated the transition from a traditional cloud data platform to an “AI Data Cloud.” Based on my expectations and the actual prints, the main takeaway is that Snowflake didn’t just meet a very high bar; they shifted the narrative from “Will AI disrupt Snowflake’s business model?” to “Snowflake is successfully monetizing the enterprise AI revolution.” Here is a look at how my expectations stacked up against their reported results and the corresponding management commentary.
In hindsight, my product revenue expectation of $1,239M might have been a bit aggressive, but Snowflake’s actual $1,227M (30% YoY growth) was still a milestone and that’s because this was the first Q4 in 5 years, where sequential net new revenue didn’t drop seasonally 20-40% QoQ, but was essentially flat. This demonstrates that underlying consumption trends are accelerating and are no longer hindered by the cloud optimization headwinds that plagued software over the last two years.
Looking ahead to the 1Q27 product revenue mid-point guide of $1264.5M (27% YoY), I am now penciling in ~$1300M of product revenue (30% YoY), which I think is well-supported by secondary forward looking growth metrics and management’s commentary. CFO Brian Robins noted that this guidance assumes “continued strength in our core business and further growth in AI workloads”. With Snowflake’s consistent history of beat-and-raise quarters, implying a 2.8% beat (after this Q’s 2.3% beat and Q3’s 2.7% beat) and a stabilization around 30% YoY revenue growth in 1Q27 feels like the right baseline.
In my last Snowflake recap, I identified RPO and cRPO as the metrics to watch, and Snowflake delivered: RPO reaching $9.77B (up 42.3% YoY) marks the second consecutive quarter of accelerating growth. The cRPO growth of 36.4% YoY (up from 32% last quarter) provides great visibility into FY27.
The Q4 earnings call provided great color on why this RPO metric exploded: Snowflake signed the largest deal in the company’s history, valued at over $400 million with a large financial services customer. Furthermore, they closed seven nine-figure contracts in Q4 alone (compared to just two in the same period last year). This isn’t just standard expansion; it is a signal that Global 2000 enterprises (of which they added 15 in Q4, up from 6 in Q3, resulting now in 790) are making multi-year, platform-level commitments to Snowflake as the main layer for their GenAI strategies. I think a cRPO acceleration of this magnitude sets a reasonable target of 30% YoY product revenue growth for FY27.
Speaking about customers, my targets for customer growth were handily beaten. Adding 740 net new customers (up 46% YoY in net additions) to cross the 13300 mark shows that the top-of-funnel motion is firing on all cylinders. In fact, customer YoY growth accelerated for the forth consecutive quarter: 4Q25: 17.2% → 1Q26: 17.5% → 18.0% → 20.0% → 21.2%. Also, the upmarket movement is impressive: $1M+ customers grew to 733, and management revealed they now have a record 56 customers spending over $10 million annually (a 56% YoY increase).
The narrative update here is all about AI product velocity. CEO Sridhar Ramaswamy highlighted that “Snowflake Intelligence has scaled from a nascent offering to an essential capability for over 2500 accounts, almost doubling quarter-over-quarter” in just 3 months. Overall AI adoption has now reached over 9100 accounts. Features like Cortex Code (now in general availability) are transforming how developers interact with the data cloud. On the call, analysts pressed on whether this AI proliferation would lead to “sticker shock” for customers. Ramaswamy addressed this head-on, noting that while consumption is expanding rapidly, they are introducing predictable pricing features (like per-user caps for Intelligence) to ensure long-term customer trust rather than short-term spikes. Nevertheless, I think this shows the potential for a consumption explosion as consumption will shift away from human-centric work loads to AI agents.
On the profitability front, Snowflake continues to prove its operational leverage. While net margin slightly dropped to 9.2% from 11.4% last year, the core operational efficiency is clear: Operating margin for the fiscal year landed at 10.5% (an expansion of over 4% YoY), and they achieved a fiscal year FCF margin of 25.5%, bringing in an impressive free cash flow of $782M in Q4 (a 61% FCF margin in Q4). In other words, In Q4, operating income grew 50% YoY and free cash flow grew 85% YoY!! And FY26 operating income grew 111% YoY while FY26 free cash flow grew 27% YoY.
Two narrative updates regarding financials emerged from the call. First, stock-based compensation (SBC) is finally being turned down, dropping from 41% of revenue in FY25 to 34% in FY26, with a firm guide to drop to 27% in FY27. Second, their recent $600M acquisition of Observe brings Snowflake into the $50B IT operations and observability market. While management warned that the Observe integration will create a ~150 basis point headwind to FCF margins in FY27 (guiding FY27 FCF margin to 23%), I think the strategic upside of owning the observability layer for AI workloads makes this a highly accretive move in the long run.
Ultimately, Q4 was a strong quarter that saw 30% revenue growth backed by a historic 42% surge in RPO. Driven by Sridhar Ramaswamy’s accelerated product velocity (since he took over), Snowflake is proving that AI isn’t cannibalizing their core consumption model - it is turbocharging it.
Overview of how Snowflake performed versus my prior expectations:
- Product revenue expectation: $1239M (7% QoQ, 31.3% YoY), implying a 3.5% beat; $80M net new product revenue.
→ Product revenue was $1227M (5.9% QoQ, 30.0% YoY), a 2.4% beat. $68M net new product revenue. While Snowflake missed my revenue expectation it is worth pointing out that this was the first Q4 in 5 years, where sequential net new revenue didn’t drop 20-40% QoQ, but was essentially flat. Maybe this just goes to show how high my expectations were. - Q1 new product revenue guide: $1276M (3% QoQ, 28% YoY) which I would interpret as $1313M (6% QoQ, 32% YoY), implying a 3% beat and that revenue growth will be roughly stable YoY.
→ New Q1 product revenue guide: $1264.5M (3.1% QoQ, 26.9% YoY), which I now interpret as $1300M (6% QoQ, 30.4% YoY), implying a 2.8% beat. - NRR around 125%.
→ NRR was stable at 125%. - I would like to see RPO around $9.85b, corresponding to 43.5% YoY growth and cRPO around $4.5b, corresponding to 37.5% YoY growth; note: RPO/cRPO is THE metric to watch in Q4: If RPO grows 25% QoQ and 43.5% YoY, then I expect cRPO to jump to 37.5%, which is my new expectation, but RPO might grow closer to 4Q25’s sequential growth of 20% (implying cRPO YoY growth staying at 32%) or 4Q24’s sequential growth of 40% (implying an insane 54% cRPO YoY growth), which is a wide range of potential outcomes that will be very telling for FY27’s revenue growth.
→ RPO was $9.77b, growing 42.3% YoY and cRPO was $4.5b, growing 36.4% YoY (up from 32.0% last quarter) and 18.8% QoQ. This supports the assumption that Snowflake will grow product revenue at 30% YoY in FY27 or maybe slightly above 30%. That’s because cRPO growth averaged 31% YoY growth in FY25 and averaged 33% YoY growth in FY26. With FY26 YoY product revenue growth of 29%, I’d expect YoY product revenue growth to go up by the delta RPO growth increase of ~2%, which would bring FY27 product revenue growth to 30-31% YoY. - I would like to see total customer growth around 4.5% QoQ (~568 adds) and $1M+ customer growth around 6% QoQ (~41 adds).
→ total customers grew 5.9% QoQ to over 13000 and adding 740 new customers. The net customer adds are up from 616 in Q3 and 506 last Q4. $1M+ customers grew 6.7% QoQ to 733, adding 46 large customers sequentially. I was also impressed by the 15 net adds in Global 2000 customers, of which they now have 790, up from 775 last quarter and 750 last Q4. - I would like to see stable edge customer growth very roughly around 6% QoQ (~310 adds) and market place listings growth very roughly around 3% (~106 adds) and AI adoption to reach about 63% of customers.
→ stable edge customer growth was 3.3% QoQ (170 net adds), market place listings grew 3.7% QoQ (132 net adds) and AI adoption reached 68% of customers. - I would like to see continued strength in profitability margins, with OM ~12%, NM ~14%, FCFM ~60%. Note: They’ll have to earn about $777M in FCF (60% margin for Q4) in order to fulfill their guided, adjusted, 25% FCF margin for FY26.
→ Operating margin was 10.8%, just as last Q, but up from 9.4% last year. Net margin was 9.2% and down from 10.8% last Q and 11.4% last year. Adjusted FCF margin was 60.9%, up from 42.9% last year, resulting in FY26 FCF margin of 25.5%. Noteworthy: In Q4, operating income grew 50% YoY! And FY26 operating income grew 111% YoY. In Q4, free cash flow grew 85% YoY! And FY26 free cash flow grew 27% YoY. - Thoughts from previous quarter: Snowflake’s 3Q26 earnings recap.
Zscaler:
Key insights:
Mixed bag quarter with:
- strong ARR & revenue growth and solid guide,
- Weak RPO growth,
- Mehh customer growth,
- Not reported NRR metrics,
- Good operating margin, but weak fcf margin.
Zscaler reported Fiscal Q2 2026 on 02/26/2026. Revenue came in at $815.8M (3.5% QoQ, 25.9% YoY), just edging past my expectation of $813M. This represents a 2.2% beat, which is perfectly in line with their historical cadence of ~2% beats we’ve seen over the last several quarters. Looking ahead, the Q3 new revenue guide of $835M (2.4% QoQ, 23.2% YoY) is solid. Assuming they maintain that standard 2% beat, I now interpret this guide as an actual target of around $852M (4.4% QoQ, 25.6% YoY), which implies their YoY growth rate is holding steady rather than decelerating significantly.
While the top line was reassuring, digging into the secondary metrics reveals exactly what I suspected last quarter regarding their recent acquisitions. In Q1, we saw a massive, inorganic spike in customer growth due to Red Canary and SPLXAI. This quarter, the hangover is apparent. Net new $100k+ ARR customers fell back to earth with 132 adds (3.5% QoQ growth), a steep drop from the 260 added in Q1, and closely mirroring the 127 added in last year’s Q2. Similarly, $1M+ ARR customers grew by 30, down from 34 in Q1 and slightly below last Q2’s 35 adds. This essentially confirms that organic enterprise growth isn’t accelerating; the Q1 numbers were just juiced by the acquired customer bases. On the Q2 call, management continued to cite the same elongated sales cycles and customer cautiousness we heard about in Q1, leaning heavily on the ZFlex program to drive cross-selling and larger deal sizes within their existing, newly expanded base.
The CFO’s new focal point, ARR, landed at $3.359b (4.8% QoQ, 25.2% YoY), which was virtually a dead center bullseye on my $3.36b expectation. RPO grew a modest 2.0% QoQ to $6.05b, while cRPO reached $2.84b (25.8% YoY), tracking very closely with the 25.9% revenue growth. However, there is a new, glaring omission this quarter: as far as I can tell, NRR was not provided. After the company abruptly stopped reporting billings last quarter - a move I already flagged as a significant yellow flag - dropping NRR from the headline metrics is frustrating. It’s highly probable that integrating the two significant acquisitions has made the NRR calculation noisy or unfavorable in the short term, but hiding a key retention and expansion metric only adds to the opacity of their organic performance.
From a profitability standpoint, Zscaler continues to demonstrate strong operational discipline, though with a slight mixed bag this quarter. Operating income came in at $181M, achieving a 22.2% margin and slightly beating my $179M expectation. However, free cash flow was a bit lighter than I wanted to see, coming in at $169M. That represents a 20.7% FCF margin, which is a noticeable contraction from the 22.1% they posted in Q2 of last year, and an expected drop from the 52.4% high-water mark we saw in Q1. While Q1 is historically their strongest cash collection quarter, the YoY margin contraction in Q2 warrants watching closely as they continue to digest their recent buyouts.
Overview of how Zscaler performed versus my prior expectations:
- Revenue expectation: $813M (3.2% QoQ, 25.5% YoY), implying a 2% beat.
→ $815.8M (3.5% QoQ, 25.9% YoY), a 2.2% beat. - Q3 new revenue guide: $829M (2% QoQ, 22% YoY) which I would interpret as $846M (4% QoQ, 25% YoY) expecting slight YoY deceleration.
→ $835M (2.4% QoQ, 23.2% YoY), which I now interpret as $852M (4.4% QoQ, 25.6% YoY), implying a 2.0% beat. - I would like to see ARR of around $3.36b (5% QoQ, 25.4% YoY).
→ ARR was $3.359b (4.8% QoQ, 25.2% YoY). - I would like to see RPO growth of around 5% QoQ (to $6.23b) and cRPO of about $2.9b (29.5% YoY).
→ RPO grew 2.0% QoQ to $6.05b and cRPO reached $2.84b growing 25.8% YoY similar to 25.9% revenue growth this quarter. - I would like to see NRR at 115%.
→ NRR was not given this quarter as far as I can tell (maybe due to two significant acquisitions complicating this calculation). - I would like to see >100k ARR customer growth around 4% QoQ (~150 net adds).
→ $100k+ ARR customers grew 3.5% QoQ adding 132 new ones, close to lat Q2’s 127 adds. - I would like to see >1M ARR customer growth around 6% QoQ (~42 net adds).
→ $1M+ ARR customers grew 4.3% QoQ adding 30 new ones, slightly below last Q2’s 35 adds. - I would like to see an operating income around $179M (22.0% margin).
→ Operating margin was 22.2% with $181M operating income. - I would like to see an FCF around $186M (22.9% margin).
→ FCF was $169M, a 20.7% margin which contracted from 22.1% last Q2. - Thoughts from previous quarter: Zscaler’s 1Q26 earnings recap.
Samsara:
Key insights:
- Exceptionally strong quarter meeting or exceeding all my prior expectations besides slightly missing my $100k+ ARR customer expectation.
- Strong revenue growth (28.3% YoY) with even stronger Q1 guide pointing to slight YoY acceleration to ~29.6%, which would reverse the FY26 trend of slow deceleration.
- Very strong ARR and net new ARR growth with total ARR growing 29.6% YoY, notably outpacing Q4 revenue growth.
- RPO growth accelerated to 42% YoY, up from 39.1% in Q3 and 32.6% last Q4 and cRPO growth accelerated to 30.9% YoY, up from 29.2% in Q3.
- Initial FY27 revenue growth guide was 22%, similar to last year’s first FY26 guide of 22.7%.
- Combine the FY27 guide with ARR and cRPO acceleration shows potential for at least revenue growth durability at 28-30% for FY27.
- This would be a big reversal of FY26’s slowly decelerating revenue growth trend.
- While large customer growth was on the weak side this quarter, the acceleration of other forward looking metrics show that expansion with existing customers is going very well.
- Samsara has always been a poster child for strongly expanding profitability metrics, which - in my mind - more than offset the slow revenue growth deceleration we saw in FY26. But now, with revenue growth durability on the table they still continue expanding profitability metrics just as before; for example: Operating margins expansion FY22: -27% → FY23: -12% → FY24: 0% → FY25: 9% → FY26: 17%.
Samsara reported Fiscal Q4 2026 on 3/5/2026, and it was an exceptionally strong finish to the year that further validates their position as a solid compounder. While Q3 was about crossing the threshold into GAAP profitability, Q4 demonstrated their ability to sustain high top-line growth at scale while driving operating leverage. On the top line, they easily cleared my expectations, delivering $444.3M in revenue (6.8% QoQ, 28.3% YoY) for a solid 5.3% beat. More importantly, the new Q1 guide is incredibly robust; they guided for $455M, but using their typical beat cadence, I now interpret this as $475M, which would represent 7.0% QoQ and an acceleration to 29.6% YoY growth.
The standout metric of the quarter, however, was the delivery of $144.8M in Net New ARR. This easily cleared my $135M target and pushed total ARR to $1.89B, growing 29.6% YoY, which is notably outpacing Q4 revenue growth. This ARR strength, combined with Remaining Performance Obligations (RPO) growing to $3.766B (up a very healthy 11.5% QoQ) and cRPO growing 30.9% YoY to $1.641B, signals that Samsara’s momentum is not slowing as they approach the $2B ARR milestone. The “system of record” value proposition remains incredibly sticky, and the backlog is building at a rate that supports sustained high-20s revenue growth.
Looking under the hood at customer metrics and the evolving narrative, there is an interesting dynamic at play. They added 204 net new $100k+ ARR customers, bringing the total to 3194 (28.6% YoY). While this volume was slightly below my target of 239, the significant beat in Net New ARR alongside the steady core customer NRR of 115% tells a clear story: expansion within the existing base is doing heavy lifting. The multi-product strategy (layering tools like AI Multicam and Asset Tags onto their core offering) is working, driving larger deal sizes and platform lock-in. It is also worth noting that management has completely shifted away from reporting the $1M+ customer count and large customer NRR. I suspect this is because their broader operations tracking narrative has matured; multi-product adoption is becoming the baseline expectation rather than an outlier, making the overall $100k+ cohort and total ARR the most relevant gauges of their success.
On the profitability front, the operating leverage story remains fully intact. Operating income came in at an impressive $91.8M, corresponding to a 20.7% margin. This is a fantastic sequential improvement from the 19.2% we saw in Q3 and a major expansion from 16.1% in the prior year’s Q4. Revenue is continuing to grow significantly faster than OPEX, proving that their R&D and go-to-market motions are highly efficient. Overall, Q4 was another classic playbook quarter for Samsara, successfully balancing top-line growth with expanding free cash flow and operating margins.
Overview of how Samsara performed versus my prior expectations:
- Revenue expectation: $442M (6.3% QoQ, 27.6% YoY), implying a 4.8% beat.
→ $444.3M (6.8% QoQ, 28.3% YoY), a 5.3% beat. - Q1 new revenue guide: $445M (0.7% QoQ, 21% YoY) which I would interpret as $466M (5.5% QoQ, 27.2% YoY).
→ $455M (2.4% QoQ, 24% YoY), which I now interpret as $475M (7.0% QoQ, 29.6% YoY), implying a 4.5% beat. - I would like to see net new ARR around $135M (total ARR around $1.88b).
→ net new ARR was $144.8M and total ARR went to 1.89b growing 29.6% YoY and faster than Q4 revenue growth. - I would like to see RPO around $3.68b (9% QoQ) and cRPO around $1.75b (39.5% YoY).
→ RPO was $3.766b (11.5% QoQ) and cRPO was $1.641b (growing 30.9% YoY). - I would like to see core customer NRR around 115%.
→ core customer NRR was 115%. - I would like to see large customer NRR around 120%.
→ not given (already since 4Q25 so I guess they no longer report this figure). - I would like to see around 239 new $100k+ ARR customers (3229 total, 30% YoY) and more than 150 total $1m+ ARR customers (wasn’t reported the last couple quarters).
→ 204 net new $100k+ ARR customers (to 3194 total, 28.6% YoY). $1M+ ARR customer number wasn’t given (already since 4Q25 so I guess they no longer report this figure). - I would like to see around $59M operating income, corresponding to a 13.4% operating margin.
→ Operating income was $91.8M, corresponding to a 20.7% margin, which is up from 16.1% last Q4, so great margin expansion. - Thoughts from previous quarter: Samsara’s 3Q26 earnings recap.
Crowdstrike:
Key insights:
- Very strong quarter, meeting or exceeding all my expectations besides slightly missing on revenue growth, but accounting for the $12.5M delta to my expectation the Q1 guide is right on point.
- Yet, revenue growth accelerated for the third quarter in a row (Q1: 19.8% → Q2: 21.3% → Q3: 22.2% → Q4: 23.3% → my Q1 projection: 24.4%), showing the same trend in forward-looking ARR growth.
- Very strong initial FY27 revenue growth guide at 23.2% YoY compared to their initial FY26 guide of 21.5% YoY.
- Another standout metric was cRPO growth which accelerated again to 33.2% YoY, up from 30.9% in Q2.
- NRRs clearly recovered this FY26, meeting my Q4 target of 115%: Q1: 112%, Q2: 111%, Q3: 114% Q4: 115%, while gross retention rates remained at 97% throughout FY26.
- Operating and FCF margins made a big jump in Q4, expanding to 25% (from 21.2% last Q4) and 28.8% (from 22.7% last Q4), respectively.
Crowdstrike’s Fiscal Q4 2026 results demonstrate the platform’s underlying stickiness and show a clear recovery from the headwinds faced earlier in the year. While the headline revenue of $1305.4M landed a bit shy of my $1318M target, the underlying business velocity tells a stronger story. The top-line shortfall is easily overshadowed by a substantial beat in Net New ARR, which surged to $330 million against my $305 million expectation, driving sequential ARR growth of 6.7%, which is up from 5.7% last Q and 5.6% last Q4. Management’s commentary reinforced that the sales environment has stabilized, with enterprise customers continuing to consolidate their security architecture onto the Falcon platform rather than maintaining disjointed point solutions.
The forward-looking metrics provide excellent visibility going into Fiscal Year 2027. Total RPO reached $9b, and crucially, current RPO (cRPO) hit $4.59 billion. Seeing cRPO growth accelerate to 33.2% year-over-year, a meaningful step up from 30.9% in Q3, suggests a strong setup for future revenue recognition. This aligns nicely with the Q1 new revenue guidance, which establishes my new expectation of around $1373m and a 24.4% YoY growth rate. The narrative from the analyst Q&A heavily leaned into this booking momentum, noting that large enterprises are making significant, long-term commitments and adopting newer offerings like next-gen SIEM and cloud security modules.
Looking at the new FY guide for FY27, the cRPO context becomes especially valuable. That’s because interpreting this annual guidance is tricky; last year’s beat of their initial FY26 guide of 21.5% was a razor-thin at 0.1%, and the new FY27 target of 23.2% suggests that if they don’t find more upside, Q1’s expected 24.4% YoY growth might represent the high-water mark for growth. However, there’s a strong argument that the initial FY26 guide they gave over a year ago was intentionally conservative due to the limited visibility following the July 2024 outage. With the air clearing, the current FY27 guide might leave more room for outperformance. This makes the cRPO growth acceleration to 33.2% from 30.9% an essential indicator that gives significant upside to this new FY27 guide.
In that context, customer retention and expansion metrics were particularly reassuring, dispelling concerns about elevated churn following the July incident. Gross retention holding firm at 97% for the entire fiscal year highlights the difficulty of ripping out Crowdstrike’s infrastructure once it is embedded. Furthermore, Net Retention Rate (NRR) steadily recovered from a trough of 111% in Q2, moving to 114% in Q3, and successfully reaching 115% in Q4. This expansion is visibly supported by module adoption, with half of the customer base now utilizing 6 or more modules. Seeing 7+ and 8+ module adoption hold steady at 34% and 24% respectively shows that cross-selling motions remain fully intact.
The profitability metrics highlight excellent operational discipline. Delivering $326M in operating income, a 25.0% margin, comfortably outpaced my $319M target. Achieving this while actively lowering Sales and Marketing spend resulted in significant operational leverage: Holding total expense growth to 17.1% YoY while revenue grew 23.3% is exactly the kind of efficiency that drives durable free cash flow. Management clearly communicated their trajectory here, working methodically toward their newly established long-term operating margin goal of 28% to 32%. The combination of accelerating cRPO, recovering NRR, and expanding margins points to a business executing very efficiently as it enters the new fiscal year.
Overview of how Crowdstrike performed versus my prior expectations:
Note: I realized after Crowdstrike reported that I accidentally copy-pasted my Q3 expectation numbers for Crowdstrike in my January recap. Here is what I actually expected for Q4 and how they did versus this expectation:
- Revenue expectation: $1318M (6.8% QoQ, 24.5% YoY), implying a 1.8% beat this Q.
→ Revenue was $1305.4M (5.8% QoQ, 23.3% YoY), a 0.8% beat. - Q1 new revenue guide: $1373M (4.2% QoQ, 24% YoY) which I would interpret as $1397M (6% QoQ, 27% YoY), expecting YoY growth acceleration.
→ New revenue guide for Q1 was for $1362M (4.3% QoQ, 23.4% YoY), which I now interpret as $1373M (5.2% QoQ, 24.4% YoY), implying another 0.8% beat. - Net new ARR of around $305M (15% QoQ and ARR +6.2% QoQ).
→ Net new ARR was $330M (24.7% QoQ and ARR grew 6.7% QoQ). - I would like to see around $9.48b RPO (20% QoQ, 45.8% YoY); cRPO around $4.65b (49% of RPO, 30.9% YoY).
→ RPO was $9.00b (13.9% QoQ, 38.5% YoY); cRPO was $4.59b; 51% of RPO and growing 33.2% YoY, up from 30.9% in Q3. This cRPO growth acceleration gives great visibility into FY27. - I would like to see NRR greater or equal to 115%.
→ NRR for Q4 was 115%. They also gave numbers for the rest of the quarters in FY26: Q1: 112%, Q2: 111%, Q3: 114% Q4: 115%. - I would like to see about $319M operating income.
→ Operating income was $326M, a 25.0% margin, up from 21.2% last Q4 and up from 21.4% in Q3. Significantly cut S&M spending in Q4 and working towards their new long-term margin goal of 28% -32%. Also, noteworthy: great operational leverage, growing expenses by only 17.1% YoY, while revenue grew 23.3% YoY. - I would like to see about $286M net income.
→ Net income was $289M, a 22.1% margin, up from 19.4% last Q4 and up from 19.9% in Q3. - I would like to see no multi-product customer decline.
→ 6+ module customers grew to 50%, up from 49% and 7+ module customers as well as 8+ module customers stayed at 34% and 24%, respectively. - I would like to see gross retention close to 97%.
→ gross retention remained at 97% for the entire FY. - Thoughts from previous quarter: Crowdstrike’s 3Q26 earnings recap.
Wrap up
Another earnings season in the books. In case you missed it, here are my write-ups of Cloudflare, Datadog, AppLovin, Astera Labs and Axon.
Overall, I was happy with how our companies performed, with clear standouts of amazing quarters by Cloudflare, Nvidia, Datadog, Axon and Samsara, and strong quarters by Crowdstrike and Snowflake.
I am wishing you all a great rest of April!
Ben
Past recaps
2022: Jul 2022 | Aug 2022 | Sep 2022 | Oct 2022 | Nov 2022 | Dec 2022
2023: Jan 2023 | Feb 2023 | Mar 2023 | Apr 2023 | May 2023 | Jun 2023 | Jul 2023 | Aug 2023 | Sep 2023 | Oct 2023 | Nov 2023 | Dec 2023
2024: Jan 2024 | Feb 2024 | Mar 2024 | Apr 2024 | May 2024 | Jun 2024 | Jul 2024 | Aug 2024 | Sep 2024 | Oct 2024 | Nov 2024 | Dec 2024
2025: Jan 2025 | Feb 2025 | Mar 2025 | Apr 2025 | May 2025 | Jun 2025 | Jul 2025 | Aug 2025 | Sep 2025 | Oct 2025 | Nov 2025 | Dec 2025