Astera Labs stock jumps 16%: is AI inference the next big catalyst?

Shares of Astera Labs ALAB surged more than 16% on Tuesday after bullish Wall Street commentary and upbeat management remarks reinforced investor enthusiasm around the company’s positioning in the rapidly evolving artificial intelligence infrastructure market.

Astera Labs stock traded as high as $255.96 during the session before changing hands near $251.28 in afternoon trading.

The sharp move higher came as analysts increasingly focused on a potential shift in AI spending patterns away from model training and toward inference workloads, an area expected to place greater emphasis on networking efficiency, connectivity, and cost optimization inside large-scale AI data centers.

Evercore ISI raised its price target on Astera Labs shares to $297 from $215 while maintaining an “Outperform” rating.

The brokerage said its industry checks suggest AI inference could become a much larger focus by late 2026 as hyperscale cloud companies increasingly prioritize cost-per-token economics and total infrastructure efficiency.

AI infrastructure focus expands beyond GPUs

Astera Labs has become one of the key beneficiaries of growing investor interest in AI infrastructure components beyond graphics processing units.

While Nvidia’s GPUs remain central to AI systems, investors are increasingly focusing on the networking and interconnect technologies required to keep large AI clusters operating efficiently.

Astera Chief Executive Officer Jitendra Mohan discussed the company’s rapid growth during the JP Morgan technology conference earlier Tuesday.

According to Mohan, Astera’s revenue has expanded from roughly $65 million around the time of its March 2024 IPO to $308 million in the latest quarter, while earnings per share increased to 61 cents from approximately 10 cents.

Management has also attempted to position Astera beyond its image as a supplier of high-speed server connectivity components.

Mohan described Astera as the “Switzerland of connectivity,” emphasizing the company’s strategy of supporting both traditional merchant GPU platforms and custom AI accelerators known as ASICs.

The company’s growing role in AI infrastructure has attracted attention as hyperscale cloud providers seek ways to reduce bottlenecks and improve utilization across increasingly expensive AI hardware systems.

Scorpio products drive growth narrative

A major focus for investors has been Astera’s Scorpio product family, which includes fabric switch chips designed to move data efficiently between AI processors.

The company said its Scorpio X-Series products featuring 320 lanes are now shipping commercially.

Patrick Moorhead of Moor Insights & Strategy said the AI bottleneck has shifted “off the GPU and into the fabric.”

SemiAnalysis founder Dylan Patel similarly noted that interconnect is “where GPU utilization goes to die.”

Astera Chief Financial Officer Desmond Lynch said the Scorpio P-Series represented the company’s fastest-growing product line last year.

Lynch also highlighted Astera’s profitability profile, calling the company’s 70% gross margin “very rich for a semiconductor business.”

On May 5, Astera reported first-quarter revenue of $308.4 million, up 14% sequentially and 93% year over year.

The company also guided second-quarter revenue to between $355 million and $365 million, while adjusted earnings per share guidance came in between 68 cents and 70 cents.

Customer concentration and competition remain risks

Despite the strong rally, Astera Labs continues facing significant business concentration risks.

The company disclosed that more than 70% of its projected 2025 revenue depends on a single customer, while its top three customers account for roughly 86% of total revenue.

Astera also warned that losing key design wins, slowing AI adoption, or increased competitive pressure could materially impact results.

Broadcom and Marvell, both identified by Astera as major competitors, possess significantly larger financial and engineering resources.