Key Takeaways:
A niche semiconductor testing company with $10.3M in quarterly revenue has secured $92M in second-half bookings from hyperscalers, betting that AI chip validation demand will outpace its fading EV business.
Aehr Test Systems, a maker of stress-testing gear that catches defective chips before they leave factories, has rallied 681% over the past year as hyperscalers place record orders to validate AI accelerators at scale. The company's burn-in and test systems subject chip samples to extreme temperatures and voltages, weeding out early-life failures that could disrupt data center operations.
"The follow-on production order from our lead hyperscaler customer demonstrates the recurring nature of our AI-related test capacity needs," Chief Executive Officer Gayn Erickson said in a statement announcing the company's fiscal third-quarter results.
Aehr reported $10.3M in revenue for the quarter ended Feb. 27, down 44% from a year earlier, as its legacy electric-vehicle chip testing business slumped. But second-half bookings reached $92M, including a record $41M production order from a top hyperscaler and a follow-on order from a global networking equipment supplier to the optical transceiver market. The company also said it has "a strong pipeline of forecasted customer orders in place."
Each hyperscaler order of that magnitude can shift revenue meaningfully for a company with a small revenue base, but the stock's 61 price-to-sales ratio leaves little room for error. Common drawdowns of 10% to 20% suggest the market is still pricing in execution risk.
Why Hyperscalers Need Burn-In Testing
AI chipmakers such as Nvidia and Broadcom sell millions of accelerators per year, but a small percentage fail shortly after deployment. Hyperscalers running clusters of tens of thousands of chips accept some failure as a cost of doing business, but elevated defect rates can disrupt training runs and inference pipelines. Aehr's equipment stress-tests batch samples under extreme conditions to catch those defects early, reducing the number of faulty chips that reach data centers.
The company has been testing its technology through deals with hyperscalers for several years and has finally started converting those pilots into production orders. The $41M record order from its lead hyperscaler came less than two weeks after management flagged it as imminent, and the networking-equipment follow-on order broadens Aehr's customer base beyond the largest cloud providers.
Valuation vs. Growth Trajectory
Aehr's revenue base remains small, which amplifies the impact of each large order. The $92M in second-half bookings compares with just $10.3M in third-quarter revenue, implying a backlog that could drive a sharp revenue acceleration once those orders convert to recognized revenue. Management has positioned the company to court multiple hyperscalers, reducing reliance on any single customer.
But the stock's 61x price-to-sales multiple is extreme by any measure. Even if revenue triples from current levels, the valuation would still imply a premium to semiconductor equipment peers such as Teradyne and Advantest, which trade at 5x to 8x sales. Investors focused on near-term valuation may find the risk-reward unattractive, while those betting on sustained AI infrastructure buildout may see the order momentum as a signal of broader demand for validation throughput.
Aehr shares, trading on the Nasdaq, have experienced 10% to 20% drawdowns multiple times during the past year's rally, reflecting the volatility that comes with a high-multiple, low-revenue-base stock. The company's ability to retain its lead hyperscaler and increase that customer's order size is a positive sign, but the path to a reasonable valuation depends on converting the current backlog into sustained revenue growth.
This article is for informational purposes only and does not constitute investment advice.
