Aehr Test Systems received a "very large forecast" from a lead AI ASIC customer, with production shipments expected to begin in Q1 FY2027. The testing equipment maker projects $60-80M in bookings for the fiscal second half ending May 2026, primarily from AI wafer-level and packaged-part burn-in systems.
The company shipped multiple Sonoma system configurations supporting up to 2,000 watts per device. Q3 Sonoma orders hit $5.5M to date, exceeding the entire Q2 total of $6.2M. Aehr claims production capacity exceeds 20 systems monthly for both wafer and package-level testing.
NVIDIA's dominance with Hopper and upcoming Blackwell architectures continues to drive specialized testing infrastructure demand. Competitors like SambaNova pursue alternative architectures with Reconfigurable Dataflow Units, while the broader ecosystem addresses computational bottlenecks from multiple angles.
Samsung's HBM3e memory advances target the "memory wall" challenge limiting AI performance. Amkor expanded semiconductor packaging capacity in Arizona, reflecting supply chain shifts for advanced AI chips. Ensurge developed microbatteries for AI-enabled edge devices, extending the hardware stack to power-constrained deployments.
Aehr's Q2 FY2026 revenue fell 27% year-over-year to $9.9M, with non-GAAP gross margin dropping to 29.8% from 45.3%. The company delayed ~$2M in WaferPak shipments for gallium nitride customers from Q2 to Q3 due to high-voltage protection circuit redesigns.
The company raised $10M through an at-the-market equity offering in Q2, selling 384,000 shares. Cash position improved to $31M from $24.7M the previous quarter, with $30M remaining under a $40M ATM program.
Aehr expanded its partnership with ISE Labs and ASE for wafer-level and packaged-part testing services targeting top-tier semiconductor customers in HPC and AI applications. The Silicon Valley test lab received multiple orders for new high-power configurations.
Second-half guidance projects $25-30M revenue with a non-GAAP net loss of $0.09-$0.05 per share. The company expects minimal silicon carbide contribution, focusing bookings on AI-related burn-in testing as custom chip development accelerates across hyperscalers and AI startups.