Scorpio X volume production begins, Astera Labs growth expectations revised sharply higher

Overview

The report believes Astera Labs is upgrading from a traditional signal-conditioning supplier to a core provider of interconnect architecture for AI system scale-out. Second-quarter results and third-quarter guidance were both significantly above expectations, with Scorpio X volume production becoming the main near-term growth driver; meanwhile, UALink, Leo CXL, optical interconnect, and custom products are expected to extend the growth cycle into 2027 and beyond.

Core views

The strength and duration of near-term growth are both better than previously expected. The content value of Scorpio X per XPU is expected to exceed $1,000, and higher attach rates, switching complexity, and differentiated features should help improve pricing and customer stickiness. Aries Gen6 remains in the early stages of ramping, Gen5 inference platforms are extending the older-generation product cycle, and the company has not observed meaningful PCIe share loss. Morgan Stanley raised its 2026 and 2027 revenue growth forecasts from 80% and 28% to 123% and 55%, respectively, but emphasized that high valuation, repeatability of results, and Amazon concentration will still create substantial volatility.

Analysis framework

The research combines quarterly results and guidance variance analysis, product-by-product ramp assessment, 2026 to 2028 financial forecasts, growth-adjusted EV/Sales valuation, and bull, base, and bear case scenario analysis, focusing on assessing the impact of Scorpio X volume production, new product expansion, and customer concentration on revenue, margins, and valuation.

Methodology notes

• Financial forecastingMorgan Stanley ModelWare

  Builds the income statement and key operating metrics based on company actual data, management guidance, and Morgan Stanley forecasts.

  Unless otherwise specified, report metrics use the Morgan Stanley ModelWare framework; forecast figures are separately identified from company-reported actuals and compared with consensus expectations.

• Relative valuationEV/Sales/Growth

  Adjusts enterprise value-to-sales multiples relative to revenue growth to assess valuation reasonableness for high-growth companies.

  The base-case price target uses 2027 revenue of $2.953B, a 2025 to 2028 revenue CAGR of 62%, and 0.37x EV/Sales/Growth, corresponding to about 23x EV/Sales. This valuation is above AI peers, but the report believes faster growth and nearly 100% exposure to AI connectivity support a premium.

• Scenario analysisBull-base-bear scenarios

  Calculates the price range under different assumptions for revenue growth, product success rates, and market share.

  The bull-case price target is $490, assuming new product ramps outperform expectations and leading share is maintained; the base case is $382; the bear case is $208, reflecting weaker new product contribution, significant Aries share loss, and a Scorpio ramp below expectations.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Astera Labs Inc（ALAB.US）

  Core research target and beneficiary of the AI connectivity theme

  Strengths

  Scorpio X is first to enter volume production; the company maintains leadership in key PCIe sockets; AI connectivity exposure is close to 100%; the product portfolio is expanding toward switching, CXL, optical interconnect, and custom solutions.

  Weaknesses

  Valuation is at an elevated level, revenue is highly dependent on a small number of customers, and near-term growth is concentrated in the main Scorpio X program.

  Comparison

  The report believes its growth rate is higher than that of major networking equipment and AI peers, supporting a valuation premium, but the absolute EV/Sales level remains high.

  Risks

  Amazon concentration, Trainium4 socket competition, Aries share loss, new product delays, a slowdown in AI capital expenditure, and non-repeatability of results.

• Amazon

  Core customer expected to contribute the majority of Astera Labs' revenue next year

  Strengths

  The collaboration covers multiple products and configurations, rather than relying solely on a single Trainium socket; Astera Labs already supports Trainium 3's full switch architecture.

  Weaknesses

  Excessively high contribution from a single customer makes Astera Labs' revenue and valuation highly sensitive to Amazon platform plans.

  Comparison

  Multi-platform collaboration reduces single-program risk, but cannot eliminate customer-level concentration risk.

  Risks

  Changes in Trainium4 design choices, platform schedule adjustments, or shifts in procurement share could significantly affect Astera Labs.

• Marvell Technology Group Ltd

  Potential competitor for the Trainium4 connectivity socket

  Strengths

  Has data center connectivity and custom chip capabilities and may participate in next-generation platform competition.

  Weaknesses

  The report believes Astera Labs has advantages from its existing hardware position, Cosmos integration, and future platform collaboration.

  Comparison

  Morgan Stanley currently views Trainium4 as an existing socket that Astera Labs needs to defend, rather than a new opportunity where Marvell has the advantage.

  Risks

  If Marvell wins a key socket, it would weaken Astera Labs' growth expectations and intensify market concerns about customer concentration.

Key data

• Second-quarter revenue$392MUp 27% quarter over quarter and 105% year over year, above consensus expectations of $361M and Morgan Stanley's forecast of $362M.
• Second-quarter non-GAAP gross margin73.7%Above consensus expectations of 72.8% and Morgan Stanley's forecast of 73.1%.
• Second-quarter non-GAAP EPS$0.80Above consensus expectations of $0.69 and Morgan Stanley's forecast of $0.70.
• Third-quarter revenue guidance$540M to $560MMidpoint of $550M, implying approximately 40% quarter-over-quarter growth, significantly above consensus expectations of $417M.
• Third-quarter EPS guidance$1.19Guidance midpoint, above consensus expectations of $0.81 and Morgan Stanley's forecast of $0.73.
• 2026 revenue forecast$1.902BCorresponds to 123% year-over-year growth, compared with the previous growth forecast of 80%.
• 2027 revenue forecast$2.953BCorresponds to 55% year-over-year growth, with the previous forecast below $2B.
• 2026 non-GAAP EPS forecast$3.99Previous forecast was $2.88.
• 2027 non-GAAP EPS forecast$5.83Previous forecast was $3.70.
• Price target and scenario rangeBase case $382; bull case $490; bear case $208Relative to the reference closing price of $361.67, these correspond to price changes of approximately 5.6%, 35.5%, and -42.5%, respectively.

Impact & implications

The Scorpio X ramp raises Astera Labs' content value and strategic position in AI infrastructure simultaneously, and drives substantial upward revisions to revenue and earnings forecasts. If the company can transfer its early advantage in PCIe to UALink, the NVLink-compatible ecosystem, CXL, and optical interconnect, growth sources will expand from a single program to multi-generation product cycles. However, the base-case price target implies only about 5.6% upside versus the report's reference share price, indicating that much of the near-term positive news has already been reflected; subsequent share performance will depend more on revenue delivery, share retention, and timely new product ramps.

Risks

• Amazon is expected to contribute the majority of next year's revenue, and customer concentration may amplify the impact of changes in orders and platform timing.
• There is uncertainty around competition with Marvell for the Trainium4 socket, and AMD's future scale-out roadmap also requires continued tracking.
• Intensifying competition could lead to market share declines for Aries retimers or other key products.
• A pause in AI spending or data center investment could cause revenue growth to slow rapidly.
• If CXL servers and 1.6T port upgrades are significantly delayed, new product revenue may fall below expectations.
• The initial Scorpio X ramp is highly concentrated, and whether above-expectation results can be sustained remains debatable.
• The stock's absolute valuation is high, while the base-case price target implies limited upside relative to the reference share price.
• Morgan Stanley and the covered company may have securities holdings, investment banking business, or other commercial relationships; investors should use this report as only one factor in decision-making.

What to watch

• Whether third-quarter revenue guidance of $540M to $560M can be achieved, and the pace of the Scorpio X volume-production ramp.
• Whether Scorpio becomes the company's largest product family ahead of schedule, and whether content value per XPU can exceed $1,000.
• The proportion of Amazon-related revenue, ownership of the Trainium4 socket, and subsequent Trainium platform planning.
• Aries Gen6 ramp, continuation of Gen5 inference platforms, and changes in PCIe competitive share.
• The launch and revenue contribution of UALink switch chips, Leo CXL, optical interconnect, and custom products in 2027.
• Whether the gross margin forecasts of 73.0% in 2026 and 70.8% in 2027 can be achieved.
• AMD's future scale-out architecture and the additional content value brought by increasing complexity in copper interconnect and optical interconnect.