Aehr Test Systems projects $60-80M in bookings through May 2026 from AI chip testing equipment, driven by a major customer forecast with production shipments starting April 2027. The Santa Clara company shipped multiple configurations of its Sonoma system supporting devices up to 2,000 watts as custom AI chip development spreads across global hyperscalers and startups.
Third-quarter Sonoma orders reached $5.5M, surpassing the previous quarter's $6.2M total. Aehr claims production capacity exceeding 20 systems monthly for wafer-level and packaged-part testing, positioning itself to serve the international rush toward application-specific AI processors that challenge NVIDIA's GPU dominance.
The testing infrastructure boom mirrors worldwide semiconductor supply chain shifts. South Korea's Samsung advances HBM3e memory to address AI performance bottlenecks, while Taiwan's packaging expertise competes with Amkor's expanded Arizona facility. Europe and Asia pursue alternative architectures like SambaNova's Reconfigurable Dataflow Units, diversifying beyond American GPU architectures.
Aehr's quarterly revenue fell 27% year-over-year to $9.9M, with gross margin dropping to 29.8% from 45.3%. The company delayed $2M in shipments for gallium nitride power semiconductors due to circuit redesigns, highlighting technical challenges in high-power testing across automotive and industrial applications spanning multiple continents.
The firm raised $10M through equity sales, improving cash reserves to $31M from $24.7M. It expanded partnerships with ISE Labs and Taiwan's ASE Group for testing services targeting top-tier semiconductor customers across high-performance computing markets in Asia, North America, and Europe.
Second-half guidance projects $25-30M revenue with a non-GAAP net loss of $0.09-$0.05 per share. Aehr expects minimal contribution from silicon carbide automotive chips, focusing instead on AI burn-in testing as global chip design fragmentates across regions seeking technological sovereignty and performance advantages beyond standardized solutions.