Date: 2026-02-22 Quarter: Q4 FY25 (Dec-25) Revenue: $270.6m | YoY: +92% | QoQ: +17.5% Non-GAAP EPS: $0.58 | Non-GAAP Op Margin: 40.2% | FCF Margin: 33.1%
Let me just say this upfront: revenue of $270.6 million, up 92% year-over-year, beating guidance by $20 million — that is a preposterous result for a semiconductor company! And the market punished the stock 20%!
I want to be clear about what happened here. Astera Labs beat on revenue, beat on earnings, delivered 40% non-GAAP operating margins, generated $282 million in free cash flow for the full year, and guided to continued growth of $291 million at the midpoint for next quarter. And the market sold it off 20% because gross margins were guided to 74% instead of 76%.
I follow the money, the results. And the results here are really, really strong.
| Quarter | Revenue | YoY | QoQ $ | QoQ % |
|---|---|---|---|---|
| Q1 FY25 | $159.4m | +144% | +$18m | +13% |
| Q2 FY25 | $191.9m | +150% | +$33m | +20% |
| Q3 FY25 | $230.3m | +104% | +$38m | +20% |
| Q4 FY25 | $270.6m | +92% | +$40m | +18% |
| Q1 FY26e | $291.5m | ~+54% | +$21m | +8% |
Look at those sequential dollar adds: $18m → $33m → $38m → $40m. The absolute dollars added every quarter are growing! That is what a healthy business looks like. The YoY percentages are "decelerating" because a year ago this company was tiny — $65 million in Q1 FY24 versus $159 million in Q1 FY25. Mathematical base effects, nothing more.
The Q1 guide of $21m sequential adds is admittedly a step-down, and I'm watching that. But I've seen ALAB guide conservatively and beat by 9% on average across four consecutive quarters. The true number is probably closer to $25-30m sequential adds.
Full year FY25: $852 million, up 115% from FY24's $396 million. That is just an astounding business.
| Quarter | GAAP GM% | Non-GAAP Op Margin |
|---|---|---|
| Q1 FY25 | 74.9% | 33.7% |
| Q2 FY25 | 76.0% | 39.2% |
| Q3 FY25 | 76.4% | 41.7% |
| Q4 FY25 | 75.6% | 40.2% |
| Q1 FY26e | ~74.0% | — |
Now people are worried about gross margins declining from 76% to 74%. Let me put this in perspective. Seventy-four percent gross margins! Show me another semiconductor company — a hardware company! — growing at 92% a year with 74% gross margins and 40% operating margins. It doesn't exist.
The margin compression is coming from Scorpio and Taurus, the newer product lines, growing as a percentage of revenue. Scorpio is an intelligent fabric switch, Taurus is smart cable modules — these are hardware products, and hardware products have lower margins than pure signal conditioning chips. But here's the thing: Scorpio and Taurus are growing even faster than the core Ares business. In one year, they went from essentially nothing to 30% of revenue! That's why the margins are compressing — the fastest-growing part of the business happens to be hardware-heavy.
This is NOT a deteriorating business. This is a maturing, diversifying business.
| | Q123 | Q223 | Q323 | Q423 | Q124 | Q224 | Q324 | Q424 | Q125 | Q225 | Q325 | Q425 | | | Mar-23 | Jun-23 | Sep-23 | Dec-23 | Mar-24 | Jun-24 | Sep-24 | Dec-24 | Mar-25 | Jun-25 | Sep-25 | Dec-25 | |---|---|---|---|---|---|---|---|---|---|---|---|---| | Revenue ($m) | 17.7 | 10.7 | 36.9 | 50.5 | 65.3 | 76.9 | 113.1 | 141.1 | 159.4 | 191.9 | 230.3 | 270.6 | | YoY % | — | — | — | — | +269% | +619% | +207% | +179% | +144% | +150% | +104% | +92% | | GM% [GAAP] | 23.7% | 78.5% | 76.2% | 77.2% | 78.2% | 78.0% | 77.8% | 74.1% | 74.9% | 76.0% | 76.4% | 75.6% | | Non-GAAP OpM | -96% | -139% | +2% | +24% | +23% | +24% | +32% | +34% | +34% | +39% | +42% | +40% | | Non-GAAP NI ($m) | -15.5 | -17.3 | -0.4 | 17.6 | 14.3 | 22.2 | 40.3 | 66.5 | 59.6 | 78.0 | 88.2 | 104.8 |
Look at that non-GAAP net income trajectory: $14m → $22m → $40m → $67m → $60m → $78m → $88m → $105m. Every quarter, more money. The business is an absolute money-printing machine. FY25 free cash flow of $282 million on $852 million of revenue — a 33% FCF margin — is simply extraordinary for a hardware company.
The company filed an 8-K revealing up to 3.3 million warrant shares tied to $6.5 billion in cumulative purchases through 2033. That is $800 million per year of implied demand from Amazon alone! Now, warrants have vesting conditions and this represents a ceiling on Amazon's commitment, not a guaranteed floor. But the structure tells you something: Amazon is embedding Astera Labs deeply into its AI infrastructure. You don't negotiate an 8-year, $6.5 billion commitment with a vendor you're planning to replace.
Five product families — Ares (PCIe retimers), Taurus (smart cables), Scorpio P (fabric switches), Scorpio X (scale-up switches), and Leo (CXL memory controllers). Scorpio X initial shipments began this quarter with material ramp expected H2 2026. Leo is now deployed in production with Microsoft's Azure M-series VMs. This is not a one-trick pony.
Management said something that really caught my attention: the merchant scale-up switching market could reach $20 billion annually by 2030, and they're targeting at least half of that. That would be $10 billion in annual revenue from Scorpio alone — against current total revenue of $270 million per quarter. The TAM expansion is real and the platform positions them for it.
AWS announced Phranyon Four support for UA Link in 2027, and AMD committed to UA Link on MI 500 series. This is the open-standard alternative to Nvidia's NVLink. Every dollar of non-Nvidia AI infrastructure that uses UA Link is a potential Astera dollar. The trend toward diversified AI compute (AMD, custom ASICs, etc.) is a secular tailwind for Astera.
R&D jumped $16 million sequentially due to the Israel Design Center and XScale acquisition. CEO said: "The TAM is much bigger than we originally expected — we are increasing our investments to pursue these opportunities." I love that! When a company with 40% operating margins chooses to invest in expansion because the opportunity is larger than expected, that is exactly what I want to see.
Mike Tate moving to "strategic advisor" and Desmond Lynch coming from Rambus: I'm watching this. CFO changes during hypergrowth are always a yellow flag. Lynch has semiconductor/finance experience, which is appropriate. The Q4 results and Q1 guidance delivered under Tate's watch were strong. I'll give this one quarter to prove out before getting concerned.
| Quarter | Guide | Actual | Beat |
|---|---|---|---|
| Q1 FY25 | $151.5m | $159.4m | +5.2% |
| Q2 FY25 | $172.5m | $191.9m | +11.2% |
| Q3 FY25 | $206.5m | $230.3m | +11.7% |
| Q4 FY25 | $250.5m | $270.6m | +8.0% |
Average beat: 9%. Apply that to Q1 FY26's $291.5m midpoint and you get approximately $318 million. That would be +17% YoY and another $48m quarter-over-quarter. That would be terrific.
| Metric | Prior | Actual | Update |
|---|---|---|---|
| Revenue beat | ~9% over guide | +8.0% | Consistent — thesis intact |
| Gross margins | ~75-76% | 75.6% GAAP | In line, Q1 guide 74% is slightly concerning |
| Operating leverage | Strong | 40.2% non-GAAP OpM | Confirmed — exemplary |
| Scorpio contribution | Growing | 30%+ of revenue | Exceeded prior expectations |
| GAAP profitability | Improving | $219m FY25 net income | Better than expected |
| FCF | Positive | $282m, 33% margin | Outstanding for hardware company |
Atlas flagged this correctly: ALAB at ~32x TTM revenue versus CRDO at ~22x with 3x the growth rate. That's a real tension. On raw PEG-equivalent for a hardware company, CRDO looks cheaper.
But here's my view: Astera is building a genuine connectivity platform — five product families, COSMOS software layer, multiple protocols, memory expansion. That platform breadth commands a premium that a pure-play cable company (CRDO) doesn't. The Amazon warrant, the TAM expansion to $25 billion, and the Scorpio X opportunity are worth something beyond just current-quarter multiples.
I am not going to add at these prices. The 20% pullback brought the stock back toward more reasonable levels but it's still expensive. At a 5% portfolio position, I'm happy to hold and let the Scorpio X ramp and continued revenue growth do the work. If the stock gives me another 20% correction — say it trades down to 25x TTM revenue — I would be a buyer.
Astera Labs is doing everything right. Revenue is growing at 92% year-over-year on a $270 million quarterly base. The platform is diversifying. The company is generating real cash — $282 million in free cash flow. Operating margins have expanded from 23% when they went public to 40% today. And they just locked in a decade-long relationship with Amazon for potentially $6.5 billion of cumulative purchases.
The stock got sold off because gross margins are guided to 74% in Q1 — one percent below the prior run rate — as faster-growing hardware products mix in. That is exactly the kind of short-term noise that creates long-term opportunity.
Thesis: Intact. Action: Hold. Would add on additional weakness below $150-160 (approximately 25x TTM revenue).
Best, Saul