If the rule were a single artificial intelligence (AI) stock for the remainder of 2026, the temptation is to default to the biggest, most-mentioned chipmaker. The problem is that the market has already priced in most of that thesis. The more interesting question is which company sells something every AI hyperscaler needs more of, has limited competition, and is still small enough to grow into a much bigger business?
By that test, Astera Labs (ALAB +1.91%) is the name that keeps standing out.
Image source: Getty Images.
The unglamorous bottleneck of AI infrastructure
Most investors hear "AI infrastructure" and think of GPUs. The reality on a data center floor is that the GPUs are useless unless the connectivity between them, the CPUs, and the memory works at very high speed and very low latency. That is the layer Astera Labs builds for. The company designs purpose-built semiconductors and modules for PCIe, CXL, and Ethernet connectivity in AI servers, and its product lines (Aries, Taurus, Leo, and Scorpio) sit inside the cabinets that hyperscalers are racing to deploy.
In May 2026, Astera Labs introduced the Scorpio X-Series 320-Lane Smart Fabric Switch, which the company calls the industry's largest open, memory-semantic fabric switch designed for large-scale AI clusters. For investors who are newer to the space, the relevance is straightforward. As AI clusters get bigger, the connections between accelerators become the new bottleneck, and the companies that solve that problem economically earn premium gross margins.
NASDAQ: ALAB
Key Data Points
A business hitting its stride
The first-quarter 2026 report tells the story numerically. Revenue hit a record $308.4 million, up 14% sequentially and 93% year over year, with GAAP gross margin of 76.3% and GAAP operating margin of 20.1%. Management guided second-quarter revenue to a range of $355 million to $365 million. Growth of that magnitude on already-high gross margins is unusual in semiconductors, and it reflects a customer set that essentially has no near-term substitute for PCIe 6 products.
Why this beats the obvious AI pick
There are bigger AI names with broader exposure. The reason to favor Astera Labs in a single-stock framework is product concentration. Almost everything the company sells goes into AI infrastructure, and that infrastructure is being built at a multihundred-billion-dollar scale by every major hyperscaler. The company's competitive position rests on early standardization wins in PCIe and CXL, and its design wins with major U.S. cloud customers have been disclosed in successive quarterly calls.
Risks worth taking seriously
The most important risk is customer concentration. A handful of hyperscalers drive most of the revenue, and any pause or rearchitecting at one of them creates a meaningful air pocket. Analysts following the company have flagged that gross margins are expected to decline 200 to 300 basis points in the second half of 2026 due to an unfavorable mix within the Scorpio X-Series and a higher share of hardware-heavy sales, which would compress profitability even as revenue scales.
Valuation is also a concern. The stock trades at a premium even by AI-infrastructure standards, and any slowdown in hyperscaler capex would compress the multiple quickly. Competition from Broadcom and Marvell in adjacent product categories is also real, even if the timing is on Astera's side for the current generation.
The takeaway
A single AI pick should be a business with a product almost every AI builder needs, gross margins that prove pricing power, and growth that is still accelerating. Astera Labs checks all three boxes today, with the caveat that customer concentration and margin compression are real near-term issues. For a year-long horizon, that combination of moat plus growth is hard to find elsewhere in the AI complex, even if the volatility along the way is going to be uncomfortable.
