Small caps spike

- Certain semiconductor and AI‑adjacent small caps have rallied sharply as the AI rebound picks up momentum. (x.com) - Example movers cited include AEHR and MARVELL‑related names rising roughly 50–100% in about a month. (x.com) - Traders are interpreting these moves as a rotation into second‑order beneficiaries of AI infrastructure demand. (x.com)

A clutch of smaller chip stocks has surged in April as investors push beyond Nvidia and other giants into companies that supply the hardware around artificial intelligence buildouts. (finance.yahoo.com) Aehr Test Systems, which makes equipment used to stress-test chips before they ship, reported April 7 that quarterly bookings jumped to $37.2 million, more than 3.5 times revenue, and said it expected second-half fiscal 2026 bookings at the high end of its prior $60 million-to-$80 million range. (aehr.com) Nine days later, on April 16, Aehr said it had won a record $41 million follow-on production order from its lead hyperscale customer for package-level burn-in systems used on custom AI processor application-specific integrated circuits, lifting second-half bookings above $92 million. (markets.financialcontent.com) Burn-in is the chip industry’s version of a stress test: manufacturers run semiconductors hot and hard to catch early failures before those parts go into servers. Aehr sells that test gear, so a rise in orders can signal that customers are preparing to ship more AI processors into data centers. (aehr.com) Marvell Technology sits farther up the stack, selling networking and custom-chip technology that moves data between processors inside AI systems. Marvell reported record fiscal 2026 revenue of $8.195 billion on March 5, up 42% from a year earlier, and said bookings were growing at a record pace. (investor.marvell.com) On March 31, Nvidia and Marvell said Nvidia had invested $2 billion in Marvell and would connect Marvell’s custom processors, networking gear and silicon photonics to Nvidia’s NVLink Fusion platform. The partnership tied Marvell more tightly to the systems companies are building to train and run AI models. (investor.marvell.com) That has fed trading in smaller names linked to the same spending wave. Applied Optoelectronics said in early April it had received an additional $71 million order for 800-gigabit data-center transceivers from a major hyperscale customer, taking that customer’s 800G orders since mid-March to $124 million. (simplywall.st) The move is a shift inside the market, not a new AI story. Instead of buying only the companies that design the flagship chips, traders are also buying businesses that test chips, connect racks, or supply the optical links that let AI servers talk to each other. (investor.marvell.com) The gains have not erased the risks. Aehr’s April 7 report also showed fiscal third-quarter revenue fell to $10.3 million from $18.3 million a year earlier, and the company posted a GAAP net loss of $3.2 million. (aehr.com) By late April, investors were still adding to the theme: Reuters reported April 19 that Google was in talks with Marvell to develop two new chips for AI inference, the stage when trained models answer real-world prompts. The rally in smaller chip names is tracking that same bet that AI spending is broadening from a few headline winners into the rest of the supply chain. (usnews.com)

Get your own daily briefing

Scout delivers personalized news, insights, and conversations tailored to your role and industry.

Download on the App Store

Shared from Scout - Be the smartest in the room.