Astera Labs (ALAB)

Investment Review After Q1 FY2026 Earnings (2026.05.05)

TL;DR

Astera Labs is a fabless connectivity-IC pure-play for AI data centers — Aries PCIe retimers, Scorpio fabric switches (P/X), Leo CXL, Taurus electrical engines — at the PCIe Gen 6, UALink, and NVLink Fusion convergence point. Just reported: Q1 FY2026 revenue $308M (+93% YoY), non-GAAP GM 76.4%, OM 36.2%, EPS $0.61 — strong beat. Q2 FY2026 guide $355-365M (+15-18% QoQ, ~+75% YoY), GM ~73% (200bps non-cash warrant impact) — continuing strong growth. Stock at $216 with 124x P/E(ttm), re-expanded from 74x at quarter close — extreme valuation that prices in Scorpio X ramp + protocol convergence wins.

Summary

Finding	Assessment
5-Year Stock Return	Stock currently $216, up ~315% from ~$52 IPO (Mar 2024). Only 2 years of trading. Revenue 7.4x ($116M→$853M) in 2 yrs. TTM OP -$116M → +$224M. P/E(ttm) 124x at $216 (re-expanded from 74x at $110 close).
Business Model	Fabless semiconductor: PCIe/CXL retimers (Aries), Ethernet cables (Taurus), fabric switches (Scorpio P/X), CXL memory (Leo). COSMOS software-defined arch. Only HV PCIe Gen 6 retimers/switches. 76% GM, zero debt, $1.18B cash.
Competitive Moat	NARROW PCIe Gen 6 timing advantage (12-18 months vs Broadcom/Marvell), COSMOS customization, Amazon $6.5B warrant + 2 new hyperscaler Scorpio P wins. But customer concentration (>70% one customer), temporary Gen 6 lead, well-resourced competitors.
Past 5-YR Drivers	PCIe Gen 6 leadership (Q1 Gen 6 >1/3 of revenue), Scorpio P-Series ramp (>15% FY2025 rev in year 1), Taurus 4x growth, fabless operating leverage (OM -127%→+25%). Amazon $6.5B warrant 2033.
Recent Stock Move	Stock currently $216, +96% post last quarter end from $110 close. Pre-print -36% QoQ on AI-spending fears, then +96% on Q1 beat + Scorpio X. Up 315% from IPO. Quarterly swings -55% to +155% — most volatile in our coverage.
Revenue Trend	ACCELERATING Rev YoY: 93% (Q1 2026) → 95% (Q2 2026E).
Margin Trend	EXPANDING OM%: 20% (Q1 2026) → 22% (Q2 2026E).
TTM OP Trajectory	DECELERATING TTM OP QoQ: 29% (Q1 2026) → 19% (Q2 2026E). TTM OP $224M.
Key Opportunities	Scorpio largest product line by year-end 2026; X-Series exceeds P-Series; ~$20B scale-up TAM by 2030. NVLink Fusion + KV cache custom + UALink intercept Trainium/MI500 (2027). Optical NPO 2027 / CPO 2028. Silicon $/XPU >$1,000.
Key Risks	P/E 124x prices in flawless Scorpio X ramp + protocol convergence wins. Customer concentration (one >70%, top 3 = 86%); Broadcom/Marvell will close Gen 6 gap; only 1 yr profitable; OpEx rigidity (700+ employees +60% YoY); Amazon warrant ~2pp GM impact starts Q2.
Catalysts	Q2 FY26 print (~Aug 2026, guide $355-365M); Scorpio X production ramp H2 2026; Computex demo; Leo CXL Azure GA end-of-year; KV cache offload custom 2027; UALink switch products 2027; NPO/optical fiber coupler ships 2027.
P/E(ttm) Valuation	ELEVATED at 124x (range: 74x–261x over 5 profitable quarters). Re-expanded from 74x at $110 close. Forward P/E ~50-60x on FY27E. Pricing in scale-up TAM expansion. Any Scorpio X miss or hyperscaler softening compresses sharply.

Q1 FY2026 Key Data

Revenue $308M (+93% YoY, +14% QoQ), beat guide. Non-GAAP GM 76.4% (+70bps QoQ); OM 36.2%; EPS $0.61. PCIe Gen 6 >1/3 of revenue (millions of ports shipped).
Cash $1.18B (flat QoQ); op cash $74.6M offset by acquisition spend. Zero debt. R&D $96M (+SBC). Scorpio X 320-lane announced — in-network compute + Hypercast; initial volume shipping.
Q2 guide $355-365M (+15-18% QoQ), EPS $0.68-0.70. GM ~73% (200bps non-cash warrant impact); OpEx $128-131M; tax 12%. Scorpio X-Series initial shipments + P-Series continued growth.

Key Takeaways from Earnings Call

Scorpio expected to be largest product line by year-end 2026 (vs 15% last year); X-Series exceeds P-Series. 320-lane X delivers 2x collective ops via in-network compute + Hypercast. P-Series adds 2 hyperscalers H2 2026; broader 2027.
Custom solutions becoming meaningful: NVLink Fusion + 2nd KV cache offload custom design (both 2027 revenue). Hybrid rack: PCIe/UALink XPUs talk to NVLink. Leo CXL on Azure M-series GA EOY. Silicon $/XPU >$1,000.
Optical roadmap: NPO ramps 2027 (XSCALE pluggable connectors), CPO mainstream 2028. Ultra-high-precision fiber coupler ships volume 2027. UALink switches intercept Amazon Trainium 4 + AMD MI500 in 2027. TAM expansion across copper + optical.

Key Segments and Revenue Mix (FY2025)

Product Line	Revenue Est.	% of Total	Key Details
Aries (PCIe/CXL Retimers)	~$400M+	~47%	Only PCIe Gen 6 retimers in HV production. +70% YoY. Core product for custom AI accelerator platforms.
Taurus (Ethernet Cable Modules)	~$200M+	~24%	400G smart cable modules for scale-out AI. +4x YoY. 800G transition next catalyst.
Scorpio (Fabric Switches)	~$128M+	>15%	P-Series: PCIe Gen 6 scale-out switch (launched Q2 2025). X-Series: scale-up switch (pre-production Q4). Fastest-growing line.
Leo (CXL Memory Controllers)	Small	<5%	Early stage. Microsoft Azure first CXL-attached memory deployment. Meaningful ramp expected 2H 2026.

Business Model

SUPPLIERS

TSMC (advanced node foundry)
OSAT partners (packaging/test)
Xscape Photonics (optical IP, acquired)
X-Scale (Israel design center, acquired)

Wafers / IP →

ASTERA LABS

PCIe/CXL retimer IC design (Aries)
Ethernet smart cable modules (Taurus)
Fabric switches (Scorpio P & X)
COSMOS software suite
Fabless: $853M rev, 76% GM, zero debt

Retimers / Switches / Cables → $ Revenue

CUSTOMERS

Amazon (>70% of FY2025 rev; $6.5B warrant)
Microsoft (Azure CXL, Scorpio)
Google (AI infrastructure)
Top 3 = ~86% of revenue
Cloud/AI server OEMs

Moat Assessment: NARROW

PCIe Gen 6 timing advantage real but temporary: 12-18 mo ahead of Broadcom/Marvell. Hyperscalers can't wait — drives design-win momentum + dollar content >$1,000/XPU. Broadcom bundling + Marvell CXL will close the gap. Extending to Gen 7.
COSMOS software + co-development create meaningful switching costs. Hyperscalers integrating COSMOS face months of re-integration to switch. Amazon $6.5B warrant through 2033 = depth. Concentration flip side: Amazon = >70% of revenue.
Scorpio X + UALink + NVLink Fusion = moat-widening vector but unproven at scale. If ALAB captures merchant silicon scale-up ($20B TAM 2030), structurally important. Protocol convergence puts ALAB at choke point. NVIDIA/Broadcom could dominate.

Stock vs. Earnings — 5-Year Performance

ALAB IPO'd Mar 2024 at ~$52; in 2 yrs of trading, stock $52 → $199 peak → $110 trough → $216 today (+315% from IPO). Revenue grew $116M (FY2023) → $853M (FY2025) — 7.4x in 2 yrs. GAAP swung -$116M loss (FY2024 IPO SBC) → +$173M profit (FY2025).
Revenue growth +93% YoY Q1 FY26 on hyperscaler PCIe/CXL adoption for AI infrastructure. Aries (+70% YoY), Taurus (+4x YoY), Scorpio (>15% of FY25 rev in yr 1). Only HV vendor of PCIe Gen 6 retimers + switches — critical Gen 5 → 6 advantage.
R40 >100% for all reported quarters since IPO — among highest in our coverage. 92-115% revenue growth + 20-29% OM. Comparable to Credo's hypergrowth. Even as growth normalizes, R40 stays >40% indefinitely.

Key Drivers of Operating Profit Growth (Since IPO)

PCIe Gen 6 leadership: only HV vendor shipping Gen 6 retimers + switches; Q1 Gen 6 >1/3 of revenue. Gen 6 doubles Gen 5 bandwidth — required for next-gen XPUs (Trainium, TPU). Multi-year cycle drives 70%+ Aries YoY. Now extending to Gen 7.
Scorpio = new higher-ASP category with "anchor socket" dynamics; largest product line by year-end 2026. P-Series >15% of FY25 rev in yr 1; X-Series 320-lane just announced (in-network compute + Hypercast). Each switch win pulls retimers + cables.
COSMOS software-defined architecture enables customization without unique tapeouts. Same silicon configured post-tape-out per hyperscaler topology. Server-grade RAS + telemetry create stickiness commodity IC suppliers can't match.

Quarterly Key Metrics

	Q1 2024	Q2 2024	Q3 2024	Q4 2024	Q1 2025	Q2 2025	Q3 2025	Q4 2025	Q1 2026	Q2 2026E
Quarter End	Mar 2024	Jun 2024	Sep 2024	Dec 2024	Mar 2025	Jun 2025	Sep 2025	Dec 2025	Mar 2026	Jun 2026
Stock Price	$74	$61	$54	$136	$61	$91	$199	$171	$110	$216
Stock QoQ	—	-18%	-12%	+155%	-55%	+50%	+118%	-14%	-36%	+96%
Rev YoY	270%	619%	206%	179%	144%	150%	104%	92%	93%	95%
OM%	-127%	-32%	-8%	0%	7%	21%	24%	25%	20%	22%
R40 (Rev YoY + OM%)	142%	587%	198%	179%	151%	170%	128%	117%	113%	117%
TTM OP ($M)	$-94	$-100	$-107	$-116	$-22	$42	$107	$173	$224	$267
TTM OP QoQ	+217%	+7%	+7%	+8%	-81%	-294%	+152%	+63%	+29%	+19%
FCF/OP ttm	12%	-44%	-85%	-88%	-496%	497%	215%	162%	153%	129%
Debt/EBITDA	0.0x	0.0x	0.0x	0.0x	0.0x	0.0x	0.0x	0.0x	0.0x	0.0x
P/E ttm	-39x	-105x	-90x	-291x	261x	162x	181x	141x	74x	124x

Recent Stock Performance & Drivers

Stock currently $216, vs $110 (Q1 FY26 close) and ~$52 IPO — +96% post last quarter end rip. -36% QoQ pre-print on AI-spending fears, then +96% on Q1 beat + Scorpio X. Up 315% from IPO. Quarterly swings -55% to +155% — most volatile in coverage.
TTM OP swung -$116M → +$224M in 6 quarters on fabless operating leverage + 115% FY2025 revenue growth. GAAP OM swung -127% (Q1 FY2024, IPO SBC) → 20% (Q1 FY2026 actual). FCF $282M+ FY2025. Zero debt; cash $1.18B = max financial flexibility.
P/E(ttm) re-expanded 74x (at $110 close) → 124x (at $216) as TTM OP keeps scaling. P/E compressed from 261x peak. R40 +113% Q1 actual; +117% Q2E. Scorecard 4/6 Q1 (R40, TTM OP>10%, FCF/OP, D/E all pass) — projects 5/6 Q2.

Opportunities

Scorpio = largest product line by year-end 2026; X-Series exceeds P-Series. X 320-lane shipping; production H2 2026. In-network compute + Hypercast = 2x collective ops. P-Series adds 2 hyperscalers H2; broader 2027. $20B scale-up TAM.
Custom solutions: NVLink Fusion + 2nd KV cache offload (2027 rev) + UALink intercept of Amazon Trainium + AMD MI500 (2027). Hybrid rack: PCIe/UALink XPUs talk to NVLink. Leo CXL Azure M-series GA EOY. Silicon $/XPU >$1,000.
Optical roadmap: NPO ramps 2027 (XSCALE connectors); CPO mainstream 2028. Ultra-high-precision fiber coupler ships volume 2027 to large AI platform provider. Israel Design Center integrated. End-to-end (copper + NPO + CPO) on same Scorpio platform.

Challenges

Valuation risk: P/E 124x (re-expanded from 74x at $110 close) prices in flawless Scorpio X ramp. Pre-print P/E 74x. Scorpio X qualification slip, Amazon design loss, or AI capex pause = sharp compression. -36% QoQ pre-print shows how fast it unwinds.
Customer concentration most extreme in coverage: one >70%, top 3 = 86%. Exceeds Credo. Demand cut / design loss / in-house Amazon silicon = devastating. $6.5B Amazon warrant locks 7+ yrs but risk persists. 2 new P-Series wins = first step.
Broadcom + Marvell will close PCIe Gen 6 gap; limited operating history (2 yrs public, 1 yr profitable, no downcycle). 12-18 mo Gen 6 lead is temporary. Broadcom bundling + Marvell CXL pressure pricing. 700+ employees +60% YoY = OpEx rigidity.

Next 12 Months

Q2 FY26 guide $355-365M (+15-18% QoQ); GM ~73% (200bps non-cash warrant); EPS $0.68-0.70. Beat >$370M with GM ≥73% extends the multiple. Miss or weak Q3 commentary compresses P/E 124x → 80x as rerating cracks.
Scorpio X 320-lane production ramp H2 2026 is the key product catalyst. Initial volume shipping today; full production H2. X exceeds P by year-end. $20B scale-up TAM 2030. Computex demo + customer quals shape Q3/Q4.
Leo CXL Azure GA end-of-year + 2nd KV cache offload custom design (2027 rev). Microsoft Azure M-series in private beta. Custom Leo for KV cache @ different hyperscaler. Memory ASP "not ASP-challenged." Optical fiber coupler 2027.

Next 12-24 Months

UALink switches intercept Amazon Trainium 4 + AMD MI500 in 2027 — protocol convergence opens new merchant TAM. ALAB UALink in development; both AWS + AMD committed. NVLink Fusion bridge chips ship 2027. Each protocol broadens XPU base.
Custom solutions becoming meaningful: NVLink Fusion + KV cache offload + memory expansions. Multi-$B TAM per mgmt. Tight hyperscaler collab + foundational tech (DSP, photonic ICs, optical packaging) from XSCALE. Multi-yr ramp 2027-28.
Optical: NPO 2027 (pluggable connector + Scorpio X optical variants), CPO 2028 mainstream. Ultra-high-precision fiber coupler shipped to large AI platform provider 2027. End-to-end (copper → NPO → CPO) on Scorpio platform — Credo-style playbook.

Core Metrics Scorecard

	Q3'21	Q4'21	Q1'22	Q2'22	Q3'22	Q4'22	Q1'23	Q2'23	Q3'23	Q4'23	Q1'24	Q2'24	Q3'24	Q4'24	Q1'25	Q2'25	Q3'25	Q4'25	Q1'26	Q2'26E
Quarter End	Sep21	Dec21	Mar22	Jun22	Sep22	Dec22	Mar23	Jun23	Sep23	Dec23	Mar24	Jun24	Sep24	Dec24	Mar25	Jun25	Sep25	Dec25	Mar26	Jun26
R40 > 40%									✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅
R40 accelerate									✅	✅		✅				✅				✅
TTM OP QoQ > 10%				✅	✅	✅	✅				✅						✅	✅	✅	✅
TTM OP QoQ accelerate						✅			✅		✅			✅			✅
TTM FCF/OP > 60%			✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	❌	✅	✅	✅	✅	✅
Debt/EBITDA < 3.5			✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅
Total passes	—	—	—	—	—	—	—	2	5	4	5	4	3	4	2	4	5	4	4	5
Stock QoQ	—	—	—	—	—	—	—	—	—	—	—	-18%	-12%	+155%	-55%	+50%	+118%	-14%	-36%	+96%
P/E ttm	—	—	N/M	N/M	N/M	N/M	N/M	N/M	N/M	N/M	-39x	-105x	-90x	-291x	261x	162x	181x	141x	74x	124x

ALAB scorecard 4/6 Q1 FY26, projects 5/6 Q2 FY26E — high-quality hypergrowth signal. R40 >40% passed every quarter since IPO. FCF/OP, D/EBITDA pass nearly always (zero debt). TTM OP QoQ >10% now consistently passing as TTM OP scales. R40 +113% Q1 (+117% Q2E) — among highest in coverage.
Stock volatility (-55% to +155% quarterly) makes scorecard timing impossible — +96% post last quarter end is the latest example. -36% QoQ pre-print (4 passes), +96% post last quarter end on beat + Scorpio X (5 Q2E). Stock moves on capex + competition more than scorecard.
R40 >40% + zero debt + 76% GM = uniquely clean financial profile in coverage. Scorecard stays strong even when stock craters. $1.18B cash + zero debt supports M&A (XSCALE, aiXscale) + R&D without financial risk. R40 staying >40% as growth normalizes compresses P/E naturally.

Chart: R40 vs Stock Price (Since 2020)

Chart: Scorecard Total Passes vs Stock Price (Since 2020)

Source: SEC EDGAR annual and quarterly filings, company earnings conference calls and presentations.

Disclaimer: This report is for educational purposes only. NO investment advice.