Can AEHR Turn AI Processor Tests Into Long-Term Revenue Streams?

Aehr Test Systems received a $22 million follow-on order in August 2026 from its lead AI processor customer, following a $14 million February order, while also securing a record $41 million hyperscaler AI-ASIC burn-in order in April. The company expects fiscal 2027 revenue of $130 million-$150 million, or 2.6x-3.0x fiscal 2026 revenue, as AI-chip production ramps drive demand for its wafer-level burn-in systems and consumables. AEHR shares have risen 111% over six months, but trade at a premium 17.51x forward P/S versus 5.54x for the industry; fiscal 2027 consensus EPS is $0.70 versus $0.03 a year earlier.

Analysis

AEHR’s investment case has shifted from technology validation to conversion risk: the relevant question is whether its installed base produces a consumables-and-expansion flywheel before customers internalize the process or diversify suppliers. Wafer-level burn-in addresses a high-value failure-cost problem in leading-edge AI devices, but system revenue remains inherently lumpy and customer concentration can turn a delayed qualification or capex pause into a material quarterly miss. The premium valuation leaves little room for a merely successful shipment cycle; investors need evidence that WaferPak/contact revenue is rising as a share of sales and that gross margin holds through customized deployments.

Over the next 1-3 months, shipment acceptance, backlog conversion, and identification of additional production customers are the catalysts that can support estimates. The key downside trigger is not weaker AI demand broadly, but a customer choosing conventional package-level test, TER/FORM-adjacent workflows, or an in-house solution for part of the flow; this would impair AEHR’s assumed attach rate and compress the revenue multiple quickly. TER and FORM offer cleaner diversified ways to express rising test intensity across compute, memory, advanced packaging, and photonics, albeit with less torque.

Contrarian view: the market may be extrapolating AI processor unit growth into test-equipment demand one-for-one. Better yields, longer equipment utilization, and concentrated hyperscaler ASIC programs can make test spend step-function rather than linear. Conversely, if reliability requirements rise faster than yields improve, AEHR’s niche could prove scarcer than consensus assumes—but that requires independent confirmation through recurring consumables, not management targets.

Key Decisions for Investors

• Do not chase AEHR after its sharp rerating; maintain a watchlist entry only after the next earnings release confirms shipment acceptance, backlog conversion, and recurring WaferPak/contactor revenue. A 15-20% post-results pullback with unchanged FY27 revenue outlook offers a more favorable entry than paying for unverified customer expansion.
• For a diversified 6-12 month AI-test exposure, favor long TER and/or FORM over AEHR. TER provides broader compute and memory test participation; FORM captures advanced packaging/HBM complexity. Use AEHR only as a smaller satellite position because its customer and qualification risk is materially higher.
• Consider a 3-6 month pair: long TER or FORM / short AEHR only if AEHR’s next report shows delayed deliveries, weaker backlog, or no incremental production-customer disclosure. The thesis is multiple compression at AEHR versus steadier estimate revisions at diversified peers; cover if AEHR demonstrates consumables growth and a second customer entering volume production.
• Set a hard thesis-review trigger on any reduction in FY27 revenue guidance, gross-margin pressure from custom systems, or evidence that a lead customer’s AI-program capex is slipping. Any of these would challenge the assumed recurring-revenue transition and could justify rapid downside repricing.