Aehr Systems stock pops

Aehr Test Systems didn’t jump because it suddenly became a household name. It jumped because investors started treating a chip-testing supplier like a bottleneck business in the artificial intelligence buildout, just days after the company reported $37.2 million in quarterly bookings and said second-half fiscal 2026 bookings should land at the high end of its $60 million to $80 million range. (aehr.com) Aehr sells machines that do “burn-in,” which is the chip industry’s version of running a new car hard before handing it to the customer. Its systems heat and power semiconductors long enough to catch early failures before those parts end up inside servers, cars, or networking gear. (aehr.com) That step is getting more valuable as artificial intelligence chips get hotter and more power-hungry. Aehr said in January that its Sonoma packaged-part systems were being ordered for next-generation central processing units, graphics processing units, and networking processors that can reach up to 2000 watts per device. (aehr.com) The market is reacting to one specific idea: if more artificial intelligence chips need longer validation before deployment, the companies selling the validation gear can grow faster than the chipmakers’ unit volumes alone would suggest. Aehr’s own April 7 release said demand was strong across both wafer-level burn-in and package-level burn-in, which means customers are paying to test chips both before and after packaging. (aehr.com) Wafer-level burn-in is the cheaper, earlier checkpoint. Instead of testing finished chips one by one after they are cut apart and packaged, Aehr’s FOX-XP systems can stress whole wafers in parallel, which lets manufacturers weed out bad dies before they become expensive finished parts. (aehr.com) That matters because artificial intelligence processors are no longer just small silicon squares. Aehr said in February 2025 that its systems can test and burn in artificial intelligence processors and accelerators either at wafer level before assembly or later as singulated die and modules, giving customers a way to screen failures at multiple stages of production. (aehr.com) The company did not post a blowout profit quarter. For the fiscal third quarter ended February 27, 2026, revenue was $10.3 million, down from $18.3 million a year earlier, and Aehr reported a non-GAAP loss of $1.5 million even as bookings surged and backlog reached $38.7 million, with effective backlog at $50.9 million including orders booked after quarter end. (aehr.com) That gap between current revenue and future orders is a big reason the stock can move so violently. Investors are looking at a company with a small revenue base, a record order pipeline, and a stated expectation of a “significant near-term follow-on production order” from its lead hyperscale customer for custom artificial intelligence processors. (aehr.com) The recent order flow helps explain why traders suddenly cared. Aehr announced a $14 million order on February 26 for fully automated FOX-XP wafer-level burn-in systems from its lead artificial intelligence processor customer, then added a new silicon photonics customer on March 31 tied to hyperscale data-center optical interconnects. (aehr.com) So the stock move is less about one tweet and more about a chain of evidence. A niche equipment maker that built its reputation testing silicon carbide chips for electric vehicles is now showing bookings, customer wins, and product specs that line up with the parts of the artificial intelligence supply chain where failure is expensive and reliability testing cannot be skipped. (aehr.com)