GauchoRico | February 25, 2026 Quarter Ending: December 31, 2025 Task Type: Earnings Review Prior Position: 12.5% (shares only)
The Feb 10 8-K filed the same night as earnings told you everything. Amazon issued Astera 3.3M warrant shares in exchange for a committed purchase agreement — up to $6.5B through 2033, averaging roughly $800M/year across smart fabric switches, signal conditioning, and optical solutions. That's more than Astera's entire FY25 revenue in a single annual guarantee from a single customer.
Let me be direct about what this is: it is simultaneously the most bullish demand signal I have seen from a semiconductor company this size, and the structural cause of every margin problem we are going to be talking about for the next six quarters.
The bull reading is obvious. $6.5B over seven years is not a capex commitment you extract from Amazon unless they have architecturally embedded your products in something they intend to build at massive scale. Amazon does not sign $6.5B deals for fun. This is Scorpio and Taurus going into AWS AI infrastructure at volume. If ALAB executes on this demand, we are looking at a company doing $1.5B+ in revenue by FY27 with a genuinely defensible position inside the largest hyperscaler on earth.
The bear reading is also obvious. Those warrants create roughly 200 basis points of quarterly gross margin drag starting Q2 FY26. Management confirmed it. Combined with the natural mix shift toward Scorpio/Taurus (lower margins than Aries), you are staring at a gross margin that goes from 77% at the peak in Q1 FY25 to potentially 70-72% by the time the mix fully rebalances. That is a 500-700bp structural compression over 18 months. For a stock at 33x TTM revenue, that is not a rounding error.
My read: the Amazon deal is a real game-changer on demand. But the margin math is also real, and the market is not wrong to reprice for it.
| Metric | Q1 FY25 | Q2 FY25 | Q3 FY25 | Q4 FY25 | Q1 FY26 Guide |
|---|---|---|---|---|---|
| Revenue ($M) | $159.4 | $191.9 | $230.6 | $270.6 | $291.5 (mid) |
| YoY Growth | +144% | +150% | +104% | +92% | ~83% |
| QoQ Growth | +13% | +20% | +20% | +17% | +8% |
| GAAP Gross Margin | 74.9% | 76.0% | 76.4% | 75.6% | ~74% |
| Non-GAAP Op Margin | 33.7% | 39.2% | 41.7% | 40.2% | ~38% |
| Non-GAAP EPS | $0.33 | $0.44 | $0.49 | $0.58 | $0.53-0.54 |
| Beat vs. Guide | +5.2% | +11.2% | +11.7% | +8.0% | — |
$270.6M revenue, +92% YoY, +17% QoQ. Eight percent beat vs. guidance on a $250.5M consensus. FY25 full year at $852.5M, +115% YoY. Non-GAAP EPS of $0.58, up from $0.49 in Q3. FCF for the full year was $281.8M — 33.1% FCF margin. Cash and securities of $1.19B with zero debt.
These are strong numbers. The revenue beat streak is four-for-four. Management has guided conservatively every single quarter and delivered above. Average beat over four quarters: +9.0%.
The one ugly data point I cannot ignore: GAAP net income collapsed from $91.1M in Q3 to $45.0M in Q4, despite $40M more revenue. That is the warrant accounting plus SBC in the same quarter. Non-GAAP NI grew to $104.8M, so the operational story is intact — but that GAAP line will make institutional screens look bad, and it needs the Q1 FY26 filing to clarify what is recurring vs. one-time.
Here is the honest trajectory:
The compression is coming from two converging forces. First, Scorpio and Taurus are inherently lower-margin products than Aries retimers — the switching and optical businesses have more hardware content and more competitive pricing pressure. Second, the Amazon warrant agreement mechanically impairs revenue recognition in a way that hits gross margin by approximately 200 basis points per quarter. You cannot easily separate these two effects going forward, which means modeling gross margin becomes guesswork until we see Q2 FY26 actuals.
For a company at 33x TTM revenue, the gross margin line is the market's signal as to whether this is a software-like infrastructure play (80%+ GM, deserves 30x+) or a systems company (65-70% GM, deserves 15-20x). At 74% and declining, we are in no-man's-land on that framework. I need to see where this bottoms.
Atlas has it right when he says "not adding until GM stabilises." I agree. The question is what stabilisation looks like — and whether Scorpio X changes the equation.
Scorpio P (scale-out switching) exceeded 15% of FY25 revenue and delivered three hyperscaler design wins. That is ahead of where I expected it to be. But the real optionality in the thesis is Scorpio X — the scale-up switching product with a stated $20B+ TAM by 2030.
Scorpio X had initial shipments in Q4. Management guided material ramp for H2 FY26. That is a six-to-nine month wait from here.
If Scorpio X ramps on schedule into the back half of calendar 2026, you have a potential reacceleration story in FY27 from a very different revenue base — higher ASPs, larger clusters, more content per customer. Combined with the Amazon commitment pulling forward Scorpio/Taurus volumes, there is a credible path to $1.5B+ revenue in FY27 with expanding margins IF the product mix improves rather than worsens.
The Leo (CXL) deployment with Microsoft/Azure/SAP is real and important as a proof-of-concept, but I am treating it as pre-revenue optionality, not a near-term driver.
The reacceleration thesis is alive but requires six to nine months of patience and execution. We will see what happens with H2 FY26 Scorpio X volume.
CFO transition: Mike Tate out, Desmond Lynch (ex-Rambus) in March 2. This is a yellow flag, not a red flag. Rambus is a relevant pedigree — IP-heavy semiconductor business — but I want to see Lynch on at least one earnings call before forming a view. Q1 FY26 call will be the first test.
Customer concentration: Top customer at 33% — down from 42% a year ago. The Amazon deal will likely push this number up again as Amazon volumes scale. Watch carefully. One hyperscaler capex pause is a 30% revenue quarter at risk.
Valuation: At ~33x TTM revenue on $852.5M TTM and decelerating toward ~83% YoY, the multiple demands continued beat-and-raise execution with no margin misses. The market's -10% AH reaction to a solid beat is a signal — the bar is high and getting higher even as the growth rate normalises. For reference, CRM at comparable growth stages traded at 10-15x revenue. ALAB at 33x requires premium pricing indefinitely. I can defend that premium IF Scorpio X ramp delivers reacceleration. I cannot defend it if we see 70% GM with 50% YoY growth in a year.
I am holding the 12.5% position. I am not adding here.
The business is executing well. Revenue, FCF, product diversification, the Amazon demand signal — all real. The concern is not whether Astera is a good business. It is whether 33x TTM revenue is the right price for a business whose gross margin is heading toward 70-72% for the next 12-18 months while growth decelerates from triple-digits toward the 50-60% range.
The fat pitch entry requires one of two things: either a GM stabilisation print (two consecutive quarters holding at 72%+) that demonstrates the floor has been found, or a Scorpio X ramp signal that credibly sets up FY27 reacceleration. Neither is visible yet. The Amazon deal makes both more likely — but timing is uncertain.
What I am watching for the trim trigger: FY26 growth dropping below 50% without visible Scorpio X ramp, or gross margin breaking below 70% without a clear path back up. Either of those forces a position review.
For now: hold the position, let Scorpio X play out, watch the Q2 FY26 print after the Amazon warrant drag begins in earnest.
We will see what happens...
GR