Aehr Test Systems' Q2 2026: Contradictions Emerge on AI Market Quantification, Booking Timing, and Customer Order Predictability

Date of Call: December 17, 2025

Financials Results

• Revenue: $9.9 million, down 27% YOY from $13.5 million
• EPS: -$0.04 per diluted share (non-GAAP net loss), compared to net income of $0.02 per diluted share YOY
• Gross Margin: 29.8% (non-GAAP), compared to 45.3% YOY

Guidance:

• Revenue for second half of fiscal 2026 expected to be between $25 million and $30 million.
• Non-GAAP net loss per diluted share for second half of fiscal 2026 expected to be between negative $0.09 and negative $0.05.
• Bookings in second half of fiscal 2026 expected to be between $60 million and $80 million, based on customer forecasts.

Business Commentary:

* Bookings and Revenue Outlook: - Aehr Test Systems reported bookings of $6.2 million in Q2 fiscal 2026, down from $11.4 million in Q1. However, bookings in the first six weeks of Q3 increased by $6.5 million, raising the effective backlog to $18.3 million. - The company expects second-half fiscal 2026 revenue between $25 million and $30 million, with bookings anticipated to be between $60 million and $80 million. - The increase in bookings was primarily driven by orders for the Sonoma system from a premier Silicon Valley test lab and other AI-related engagements.

• AI and Semiconductor Test Demand:
• The company is observing significant demand in both wafer-level and packaged-part burn-in segments, particularly for AI processors.
• A key growth strategy is focused on reliability solutions for AI and data center infrastructure, leading to increased capacity and potential system demand.

• The massive explosion of AI and data center infrastructure is a primary driver for this growth.

• Segment Performance and Market Expansion:

• In wafer-level burn-in, Aehr Test Systems expanded engagements across AI processors, flash memory, silicon photonics, gallium nitride, and hard disk drives.
• The company secured new device wins for the Sonoma system, supporting high-temperature operating life qualifications for AI devices, with potential for substantial volumes.

• Market expansion into AI processors, gallium nitride power semiconductors, data storage devices, silicon photonics, and flash memory is diversifying the company's customer base and increasing its total addressable market.

• Financial Performance and Guidance:

• Revenue for Q2 fiscal 2026 was $9.9 million, down 27% from the prior year period, primarily due to lower shipments of WaferPaks.
• Non-GAAP gross margin declined to 29.8% from 45.3% a year ago, impacted by lower sales volume and a less favorable product mix.
• The company is reinstating guidance for the second half of fiscal 2026, driven by increased visibility from customer forecasts and a strong outlook for fiscal 2027.

Sentiment Analysis:

Overall Tone: Positive

• CEO states: 'Taken together, our increased visibility across multiple end markets gives us great confidence in our outlook.' Also notes: 'we're reinstating guidance in fiscal '26' and highlights 'very strong fiscal '27' prospects. 'We feel we're very well positioned' and 'we believe our bookings... will be between $60 million and $80 million.'

Q&A:

• Question from Christian Schwab (Craig-Hallum): What wasn't clear to me exactly is on the booking strength -- potential booking strength of $60 million to $80 million in the second half of this fiscal year. Is that almost entirely on the AI accelerator processor line?
  Response: The bulk is across wafer-level and packaged-part burn-in for AI processors, with some silicon carbide and silicon photonics.

• Question from Christian Schwab (Craig-Hallum): Given the AI processor bookings, can you indicate the multiyear opportunity including '27 and '28?
  Response: The AI business could be measured in hundreds of millions of dollars for Aehr Test a few years out, with significant market potential.

• Question from Christian Schwab (Craig-Hallum): How many systems do you think you're capable of manufacturing in a year for wafer level?
  Response: Capable of manufacturing up to 20 systems a month at either package or wafer level if needed.

• Question from Jonathan Dorsheimer (William Blair): On the wafer level benchmark, is the delay due to changing parameters from the customer?
  Response: The delay was due to initial customer parameters based on package-level thinking, which required adaptation for wafer-level testing, but progress is on track.

• Question from Jonathan Dorsheimer (William Blair): Are you anticipating cannibalization between package and wafer level for AI processors?
  Response: Both package and wafer-level testing will coexist for a long time; wafer-level offers yield advantages, but customers may choose based on their needs.

• Question from Maxwell Michaelis (Lake Street Capital): To get to the $80 million booking upper end, does that suggest improvement in silicon carbide or GaN?
  Response: The $80 million suggests greater volume orders primarily from wafer-level and packaged-part AI burn-in, with lesser contributions from silicon carbide, GaN, and hard disk drive.

• Question from Maxwell Michaelis (Lake Street Capital): Can you share more detail on the $5.5 million order?
  Response: The order includes a mix of AI-related Sonoma systems, burn-in modules, and high-power configurations for a test services company, important for upcoming production.

• Question from Larry Chlebina (Chlebina Capital): Would customers cut over to wafer-level burn-in midstream after a product launch?
  Response: It depends on the customer and application; wafer-level offers significant yield and cost advantages, but transition timing varies.

• Question from Larry Chlebina (Chlebina Capital): When do you expect the flash benchmark customer to place an order?
  Response: Expected to hear back in the next couple of months, with potential orders following validation and design reviews.

Contradiction Point 1

AI Market Opportunity Quantification

This is a substantial contradiction involving a change in how the long-term AI market opportunity is quantified, shifting from a relative comparison to silicon carbide to an absolute dollar range. This change provides a more specific (and significantly larger) target for investors, altering the perceived scale of the opportunity.

Are the $60M–$80M booking strength in H2 FY24 primarily driven by the AI accelerator processor line? Given the significant bookings from the AI processor market, can you provide guidance on its multi-year growth trajectory, including 2027 and 2028? What is your annual production capacity at the wafer-level? - Christian Schwab (Craig-Hallum)

20260109-2026 Q2: The AI market opportunity is large, with estimates suggesting total AI test and burn-in spend could be $8-15 billion. - Gayn Erickson(CEO)

Can you estimate the long-term AI market opportunity, like the prior silicon carbide model? - Christian David Schwab (Craig-Hallum Capital Group LLC)

2025Q4: The AI market potential is estimated to be 3 to 5 times larger than the silicon carbide market was at a similar future point. - Gayn Erickson(CEO)

Contradiction Point 2

AI Market Opportunity and Booking Timing

This is a substantial contradiction reflecting a change in the narrative around the expected timing for material bookings/orders. It moves from cautious, specific guidance on the timing of production orders to a more forward-looking statement that avoids near-term commitment, potentially indicating a shift in the perceived sales cycle duration or order pipeline confidence.

Is the $60M–$80M booking growth in H2 FY24 primarily from AI accelerators? Can you provide guidance on the AI processor business's multi-year growth (including 2027–2028)? How many wafer-level systems can you manufacture annually? - Christian Schwab (Craig-Hallum)

20260109-2026 Q2: The AI market opportunity is large, with estimates suggesting total AI test and burn-in spend could be $8-15 billion. Aehr's ASPs are higher for wafer-level systems, and the company is in a strong position... potentially seeing hundreds of millions in revenue from AI in a few years. - Gayn Erickson(CEO)

What is the expected timeframe for the AI-driven TAM to become evident compared to silicon carbide? Should we expect material orders this fiscal year? - Christian Schwab (Craig-Hallum Capital Group LLC)

2026Q1: Management is confident in the TAM growth but is not providing specific guidance on timing... The sales cycle for ramping new customers is long, so production orders are not expected until closer to the end of the current fiscal year (May 2026). - Gayn Erickson(CEO)

Contradiction Point 3

Timing of AI Customer Orders

This is a substantial contradiction regarding the expected timeline from customer evaluation to placing an order. It shifts from citing a concrete historical example (a first customer moving from evaluation to order in ~6 months) to a more generic statement about the evaluation timeline, creating uncertainty about the predictability and speed of the sales cycle for new AI customers.

Why was the wafer-level benchmark for a key AI customer delayed? Are you expecting cannibalization between package and wafer-level burn-in for AI processors and ASICs? - Jonathan Dorsheimer (William Blair)

20260109-2026 Q2: The evaluation is still on track, with data expected in the coming months. - Gayn Erickson(CEO)

For new AI clients, how long does the qualification process take to reach a final decision and order? - Unidentified Analyst (Freedom Broker)

2025Q4: The timeline from initial evaluation to order varies. The first AI customer moved from evaluation to order within about six months, which is faster than the second customer. - Gayn Erickson(CEO)

Contradiction Point 4

Nature of the Customer List Disclosure

This is a substantive contradiction concerning the transparency and definition of the company's customer base. It involves a shift in the characterization of the disclosed customer list from explicitly representing "current" customers to potentially including "prospects," which could mislead investors about the quality and near-term convertibility of the pipeline.

Is the $60M–$80M booking strength in H2 FY24 primarily from the AI accelerator processor line? Can you share expectations for AI processor business growth over the next few years (including 2027 and 2028)? What is your annual production capacity at the wafer-level? - Christian Schwab (Craig-Hallum)

20260109-2026 Q2: The bulk of the bookings is across wafer-level and packaged-part burn-in for AI processors, with some contribution from silicon carbide and silicon photonics. - Gayn Erickson(CEO)

Are the marquee customer names on the slide current/previous customers or prospective ones (e.g., new AI or silicon photonics customers)? - Christian David Schwab (Craig-Hallum Capital Group LLC)

2025Q4: The updated customer list reflects current Incal package part burn-in customers, not prospects. - Gayn Erickson(CEO)

Contradiction Point 5

Flash Memory Benchmark Customer Feedback and Order Timing

This is a substantial contradiction regarding the expected timeline for a business-critical event—the receipt of a production order from a key flash memory customer. It moves from expressing a specific expectation for customer feedback (within "a couple of months") to a more uncertain stance where the customer has taken wafers back and the order timing is uncertain. This directly impacts short-term revenue forecasting.

When do you expect an order from the customer following the pre-holiday flash memory benchmark? - Larry Chlebina (Chlebina Capital)

20260109-2026 Q2: The customer is expected to provide feedback in the next couple of months. - Gayn Erickson(CEO)

Can a customer start with package-part burn-in for a new product and later switch to wafer-level burn-in after validation? Regarding the flash memory benchmark, when do you expect the customer to place an order? - Larry Chlebina (Chlebina Capital)

2026Q2: Regarding the completed flash benchmark, the customer has taken the wafers back for further processing.... A production order is expected, but timing is uncertain. - Gayn Erickson(CEO)