ALAB — Earnings Review Q4 FY25 (Atlas)

Date: 2026-02-23 (updated from 2026-02-21 with stage outputs: quant-prep, transcript-digest, scuttlebutt) Quarter: Q4 FY25 (Dec-2025) Market cap: $28.7B | EV/TTM Rev: ~32x | Revenue growth: 92% YoY

Verdict

Astera Labs delivered a clean Q4 beat — $270.6M revenue (+92% YoY, +17.5% QoQ) against $250.5M guide, the 4th consecutive beat averaging 9%. The core business is executing. But three things dominated the post-earnings reaction and justify the ~20% stock drop: gross margin guided to 74% for Q1 (structural, not cyclical), GAAP net income collapsed from $91.1M in Q3 to $45.0M in Q4 despite $40M revenue growth, and the CFO is departing mid-hyper-growth. Scorpio+Taurus jumped from ~15% to ~30% of revenue in a single quarter — a product inflection that expands TAM but compresses margins. The Amazon $6.5B warrant provides long-term demand visibility at an unknown price point. At ~32x TTM revenue decelerating toward 37% annualized (per Q1 guide), valuation leaves no margin for error. Conviction: 3 — hold. Not adding until gross margin trajectory stabilises and Scorpio X ramp demonstrates growth reacceleration.

Qualification Gate

All six gates pass. SBC at 15.3% of revenue (declining from 18.8% a year ago) is appropriate for a post-IPO semiconductor company in hyper-growth.

Six-Factor Score

The Numbers

| | Q1_FY23 | Q2_FY23 | Q3_FY23 | Q4_FY23 | Q1_FY24 | Q2_FY24 | Q3_FY24 | Q4_FY24 | Q1_FY25 | Q2_FY25 | Q3_FY25 | Q4_FY25 | | | Mar-2023 | Jun-2023 | Sep-2023 | Dec-2023 | Mar-2024 | Jun-2024 | Sep-2024 | Dec-2024 | Mar-2025 | Jun-2025 | Sep-2025 | Dec-2025 | |---|---|---|---|---|---|---|---|---|---|---|---|---| | Revenue ($M) | 17.7 | 10.7 | 36.9 | 50.5 | 65.3 | 76.9 | 113.1 | 141.1 | 159.4 | 191.9 | 230.3 | 270.6 | | YoY % | — | — | — | — | +269% | +619% | +207% | +179% | +144% | +150% | +104% | +92% | | QoQ % | — | -40% | +245% | +37% | +29% | +18% | +47% | +25% | +13% | +20% | +20% | +18% | | GM % [GAAP] | 23.7% | 78.5% | 76.2% | 77.2% | 78.2% | 78.0% | 77.8% | 74.1% | 74.9% | 76.0% | 76.4% | 75.6% | | Op Margin [NG] | — | — | 1.9% | 24.4% | 22.7% | 24.3% | 32.4% | 34.3% | 33.7% | 39.2% | 41.7% | 40.2% | | GAAP NI ($M) | -15.5 | -17.3 | -0.4 | 17.6 | 14.3 | 22.2 | 40.3 | 66.5 | 59.6 | 78.0 | 91.1 | 45.0 | | GAAP NI Margin | — | — | neg | 35% | 22% | 29% | 36% | 47% | 37% | 41% | 40% | 17% | | NG 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 | | NG NI Margin | — | — | neg | 35% | 22% | 29% | 36% | 47% | 37% | 41% | 38% | 39% | | EPS [NG] | — | — | — | — | — | — | — | — | $0.33 | $0.44 | $0.49 | 0.58||FCF(M) | — | — | — | — | — | — | — | — | 6.0 | 133.3 | 65.9 | 76.6 |

FY25 totals: Revenue $852.5M (+115% YoY). Non-GAAP EPS $1.84. GAAP EPS $1.22. FCF $281.8M (33.1% margin). Zero debt. $1.19B cash + marketable securities.

GAAP/NG flag — Q4 divergence: Q3→Q4 GAAP NI collapsed 91.1M→45.0M despite $40M revenue growth. Non-GAAP NI rose 88.2M→104.8M. The 59.8MQ4GAAP/NGgap(vs 3M in Q3) reflects elevated SBC (~$41M at 15.3% of revenue) plus likely discrete tax items. This is not an operational deterioration — Non-GAAP tells the operating story. But the magnitude of the Q4 GAAP swing warrants monitoring in the 10-Q for the root cause.

Prior Beliefs / Updated Beliefs

Delta Assessment

The beat itself was expected — ALAB has beaten guide by an average of 9% over 4 quarters, and Q4's 8% beat was the smallest in the streak. Three surprises dominated, net negative to sentiment:

1. Gross margin guided to 74% for Q1, with Amazon warrant adding ~200bp drag from Q2. This is the most important number in the quarter. It confirms the mix shift toward Scorpio/Taurus is accelerating faster than market expected, and that Amazon negotiated significant pricing concessions. Gross margin trajectory is clearly downward; the floor is unknown.

2. Scorpio 20% Q4 target — apparent miss. Management deflected a direct analyst question on whether Scorpio hit 20% of Q4 revenue. The language suggested 15-20%. If it had cleanly hit 20%, they would have said so. Scorpio+Taurus combined at ~30% means Scorpio P alone was likely 15-18%. Not catastrophic, but it establishes that management's aspirational targets are not commitments.

3. GAAP NI collapse (91.1M→45.0M) alongside Non-GAAP growth (88.2M→104.8M). The $46M swing in GAAP net income alongside $40M revenue growth is unexplained until the 10-Q files. Most likely cause: discrete tax items and elevated stock vesting. If operational, it would show up in Non-GAAP — it didn't. Still, the opacity added to investor unease.

Leading Indicators

No SaaS-style ARR/RPO/billings — semiconductor company. Relevant proxies:

• Beat magnitude trend: +5.2% → +11.2% → +11.7% → +8.0%. Narrowing. Four-quarter beat streak is a strong track record, but the cycle high was Q3 FY25.
• Sequential dollar adds: 18→33→38→40M over FY25. Stable. Q1 FY26 guide implies +$21M QoQ — the sharpest deceleration in the year. Could be seasonal; management's track record of conservatism is the mitigating factor.
• Product mix shift as demand breadth proxy: Scorpio+Taurus at ~30% of Q4 revenue (up from ~8% Q1 FY25) signals new product adoption is accelerating across the customer base — not just Aries repurchases. Bullish for TAM; bearish for margins.
• Customer concentration: ~42% top customer → ~33% over FY25. Diversification is real and sustained. Base expanding across 10+ hyperscalers.
• OpEx investment signals: R&D up $16M QoQ, Israel Design Center expansion, 37% headcount growth (440→602). Management is spending ahead of revenues — historically bullish when conviction is high on pipeline.
• Hiring specificity: 38 open roles concentrated in AI Fabric engineering — directional signal aligned with Scorpio X ramp priorities.

Summary: Mixed signals. The product mix inflection (15%→30% Scorpio+Taurus) is the strongest positive leading indicator. Beat narrowing and Q1 QoQ guide step-down are the softest signals. No clear divergence.

Scuttlebutt Findings

• Market reaction: Stock dropped 20%+ post-earnings despite revenue beat. Consensus narrative: margin compression + Amazon margin risk + CFO departure > revenue beat + Amazon warrant. Citi cut PT 275→250 but maintained Buy. 18/23 analysts rate Buy or Strong Buy, consensus 12-month PT ~$191. (CNBC, 24/7 Wall St, scuttlebutt stage)
• Amazon warrant framing: Buy-side and sell-side viewed the $6.5B commitment as structural — not a standard cancellable PO. Demand floor is credible. The margin structure of the warrant is the unanswered question. Management refused to disclose pricing terms. (Scuttlebutt stage)
• Competitive threat — Broadcom: Broadcom AI networking revenue +170% YoY, directly in scale-up Ethernet where Scorpio competes. Bundling risk (Broadcom custom ASIC + Ethernet switch) is the primary bear thesis circulating on the street. SemiAnalysis has written on this dynamic. (SemiAnalysis, scuttlebutt stage)
• Competitive threat — CRDO/Marvell: CRDO is the direct AEC competitor vs Taurus. Marvell acquired XConn (CXL memory controller), directly adjacent to Leo. Competition intensifying at multiple product layers simultaneously. (Scuttlebutt stage)
• Employee sentiment: Glassdoor 4.3/5 (109 reviews), 87% recommend. Blind 4.2/5. Compensation and career growth rated high; work-life balance weak (3.1-3.8). Consistent with hyper-growth startup culture — no signals of cultural breakdown. (Glassdoor, Blind)
• CEO Jitendra Mohan: 20+ years in interconnects, technically credible. Paragon Intel note: "untested in downturn." He has not managed a cyclical semiconductor correction as CEO. Worth monitoring if AI capex softens. (Scuttlebutt stage, Paragon Intel)
• COSMOS software ecosystem: Customer feedback indicates genuine switching costs from toolchain integration and firmware update dependencies. Meaningful moat element beyond silicon. (Futurum)
• CFO Desmond Lynch (incoming): Ex-Rambus — smaller, IP-licensing model, different operating context. Transition during FY26 when the margin narrative is most complex is poor timing. (Scuttlebutt stage)
• Market sizing paradox: Analyst firms peg the retimer TAM at $613M (2024) → $1.02B (2029). ALAB's $852M FY25 revenue already eclipses that 2029 estimate. This confirms ALAB is not a retimer company — Scorpio, Taurus, and Leo are distinct product categories. TAM expansion thesis is real, not promotional. (Scuttlebutt stage)

Valuation Context

Growth-adjusted comparison: EV/Rev ÷ YoY growth: ALAB ~0.35x vs CRDO ~0.08x. ALAB's platform premium is justified in concept — 5 product families vs CRDO's narrower AEC focus. But the magnitude of the gap implies the market is already pricing significant platform optionality into ALAB.

3-year intrinsic framing: If ALAB delivers 40% CAGR revenue through FY28 (conservative for hyper-growth, aggressive given deceleration signals), TTM revenue reaches ~2.3B.At20xTTMrevenue(amaturegrowthmultiple), marketcap 46B — roughly 60% upside over 3 years. Acceptable, not exceptional, for this risk profile. The Scorpio X ramp is the upside scenario; gross margin deterioration to <70% is the downside scenario.

Q1 FY26 guidance: Revenue $291.5M midpoint, GM ~74%, Non-GAAP EPS $0.53-0.54. EPS guidance implies earnings growth is slowing even as revenue grows — consistent with OpEx expanding $100M run-rate. Near-term earnings compression; longer-term operating leverage depends on Scorpio X ramp.

Platform & Secular Position

• Secular trend: AI infrastructure buildout — connectivity layer (PCIe, Ethernet fabric, optical, memory) within AI clusters. Among the strongest secular tailwinds in technology. Hyperscaler capex growing 40-50% YoY in 2026.
• Platform vs. point solution: Genuine platform. Five product families: Aries (PCIe retimers/switches), Taurus (optical AECs), Scorpio P (scale-up Ethernet 100G-800G), Scorpio X (scale-up switching 1.6T+), Leo (CXL memory expansion). COSMOS software layer unifies the stack.
• TAM penetration: Current SAM ~$2.5B; ALAB at $852M TTM = ~34% current SAM penetration. Management projects 10x expansion to $25B by 2030 — implying <4% of future TAM. The merchant scale-up switching TAM alone (Scorpio X) is $20B+ by 2030.
• Protocol positioning: UA Link vs NVLink vs Scale-Up Ethernet is a live standards war. UA Link backed by AWS and AMD; NVLink dominates training. ALAB's software-defined approach maintains protocol flexibility — the right hedge in an unsettled standards landscape.
• Optical timeline: Scale-up optical products targeted ~2028. Two-year wait on a key TAM vector. CXL (Leo) is the nearer-term diversification play.

Key Risks

1. Gross margin compression is structural. As Scorpio and Taurus grow from 30% to majority of revenue, hardware content rises. Guided 74% Q1, likely trending toward 70-72% at scale. Amazon warrant adds ~200bp additional headwind Q2+. This is not cyclical.
2. Amazon concentration and pricing power. $6.5B warrant through 2033 locks in demand but Amazon negotiated the terms — likely significant volume discounts. If Amazon scales to 40-50% of revenue, margin drag becomes the dominant business narrative.
3. GAAP NI opacity. Q3→Q4 collapse (91.1M→45.0M) alongside Non-GAAP stability unexplained until 10-Q. If it recurs in Q1 FY26, it becomes a credibility and investor relations issue.
4. CFO transition timing. Desmond Lynch (ex-Rambus) enters during the most complex margin narrative in the company's public history. Risk of investor relations stumbles on the warrant economics and OpEx investment story.
5. Broadcom bundling threat. If Broadcom sells scale-up Ethernet switching tied to its custom ASIC business, ALAB's Scorpio TAM could be structurally capped at non-Broadcom-ecosystem accounts. Broadcom AI networking +170% YoY — they are not standing still.

Key Catalysts

1. Scorpio X-Series material ramp H2 2026. Management targets Scorpio X as eventually the largest product family. Any confirmation of >$50M/quarter contribution by Q3 FY26 would reaccelerate growth and validate the TAM expansion narrative.
2. Gross margin floor. If GM stabilises at 72-74% rather than continuing to decline, the bear case loses its primary pillar. Any guide for stable GMs in H2 FY26 would be a significant positive catalyst.
3. UA Link protocol wins. AWS Phranyon Four and AMD MI 500 targeted for 2027. If UA Link is publicly adopted at scale before that, ALAB's Aries+Scorpio design wins expand significantly.
4. Customer diversification below 30% top-customer. Currently ~33%. Further diversification reduces Amazon margin leverage and demonstrates the platform is not a single-customer story.
5. Leo/CXL scaling with Azure. Azure M-series deployment is live. Any announcement of broader cloud adoption (GCP, AWS) of Leo-based CXL memory opens a TAM vector worth $5B+ by 2028.

Promise Tracker

Pattern: Strong delivery record on major metrics (4 consecutive revenue beats, product milestones). One apparent miss (Scorpio 20% Q4). Deflection on the specific target rather than a miss on the broader category is the more charitable read — but management set an aspirational target they then retreated from without clean disclosure.

Analysis by Atlas. Supersedes 2026-02-21 version — incorporates quant-prep, transcript-digest, and scuttlebutt stage outputs. No position held.