Aehr Test Systems Q2 2026: Contradictions Emerge on AI Market Size, Customer Evaluations, and HBF Opportunity

Date of Call: December 17, 2025

Financials Results

• Revenue: $9.9M, down 27% YOY from $13.5M
• EPS: Negative $0.04 per diluted share, compared to net income of $0.02 per diluted share in the prior year
• Gross Margin: 29.8% (non-GAAP), compared to 45.3% a year ago

Guidance:

• Revenue for the second half of fiscal 2026 (Nov 29, 2025 - May 29, 2026) expected to be between $25 million and $30 million.
• Non-GAAP net loss per diluted share for the second half expected to be between negative $0.09 and negative $0.05.
• Bookings in the second half of fiscal 2026 expected to be between $60 million and $80 million, based on customer forecasts.
• This sets the stage for a very strong fiscal 2027.

Business Commentary:

* Revenue and Booking Outlook: - Aehr Test Systems reported second quarter revenue of $9.9 million, which was down 27% from $13.5 million in the prior year period. However, the company anticipates bookings in the second half of fiscal '26 to be between $60 million and $80 million. - The revenue decline was primarily due to lower shipments of WaferPaks, while the positive booking outlook is driven by increased visibility across multiple end markets, particularly AI and data center infrastructure.

• AI and Semiconductor Test Market Growth:
• The company is seeing significant progress in both wafer-level burn-in and packaged-part burn-in segments, with a focus on AI processors, flash memory, silicon photonics, and gallium nitride.

• This growth is attributed to the massive explosion of AI and data center infrastructure, which is driving a need for comprehensive test and burn-in solutions.

• Sonoma System Demand:

• Aehr received orders exceeding $5.5 million for its Sonoma ultra-high-power packaged-part burn-in systems in the third quarter to date, surpassing the total orders for the entire second quarter.

• This increase in demand is driven by the need for high-temperature operating life qualifications for AI devices and the system's capability to support high-volume production.

• GaN and Silicon Carbide Developments:

• There has been a delay in expected orders for gallium nitride (GaN) due to unanticipated high-voltage fault conditions, while silicon carbide demand has been pushed out to next fiscal year.

• The delays are related to design and system enhancement challenges, but the company remains well-positioned for future demand in these markets.

• Flash Memory Benchmark and Market Opportunity:

• Aehr completed a wafer-level benchmark with a global leader in NAND flash, demonstrating its ability to test flash memory wafers with higher parallelism and power than traditional methods.
• This positions the company to address the growing demand in the memory market, particularly for emerging devices like High Bandwidth Flash (HBF).

Sentiment Analysis:

Overall Tone: Positive

• Management expresses excitement about prospects: 'we made significant progress... and are very excited about our prospects moving forward', 'we believe our bookings... will be between $60M and $80M, which would set the stage for a very strong fiscal '27', 'we feel we're very well positioned', 'we're reinstating financial guidance... with great confidence in our outlook'.

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 contributions from silicon photonics and minimal silicon carbide.

• Question from Christian Schwab (Craig-Hallum): Given that such a material bookings from the AI processor market, can you give us any indication or idea? I know we've talked about the opportunity in that marketplace being bigger than silicon carbide. But let's narrow it down to kind of a multiyear time frame kind of including '27 and '28. Do you see that business after initial orders expanding meaningfully from there?
  Response: Yes, the AI business could be measured in hundreds of millions of dollars for Aehr Test a few years out, with significant potential given the large market size and Aehr's strong position.

• Question from Christian Schwab (Craig-Hallum): And then I guess one last question, if I may, and follow up on your comment about capacity. How many systems do you think you're capable of manufacturing in a year for wafer level?
  Response: The company could manufacture up to 20 systems per month (wafer and packaged-part combined) if needed, exceeding current forecasts.

• Question from Jonathan Dorsheimer (William Blair): I guess maybe just to start, on the wafer level, I think your prior comments around the timing of the benchmark, it seems like that's taken a little bit longer. And I'm just wondering, is that a function of -- is it because it's new and what you're seeing is from the customer is that they're changing parameters that's extending that out?
  Response: The delay was due to the customer initially providing test vectors based on package-level assumptions, which required adaptation for wafer-level testing; the benchmark is now on track for completion in the coming months.

• Question from Jonathan Dorsheimer (William Blair): And I just want to address the potential of cannibalization between package and wafer level. And if I read through your comments, it seems like the AI processor is what's moving along with this customer on the wafer level. You had mentioned briefly actually on the ASIC side. Do you -- are you anticipating that the ASICs basically run with package level and that AI processors are wafer level? Or are you anticipating both at wafer level?
  Response: Both package and wafer-level solutions will coexist for a long time; there is potential cannibalization, but Aehr is positioned to serve both, with wafer-level offering significant yield advantages.

• Question from Maxwell Michaelis (Lake Street Capital): First one for me, just around the bookings guide. I know you previously shared that majority of around AI. But just given the distinction between the low end and the high end, if we just take the midpoint of around $70 million, I mean, what -- to get to that $80 million, is that all basically around AI? Or does that suggest any improvement around silicon carbide or GaN?
  Response: The higher end of the bookings range includes additional volume from wafer-level burn-in AI, with contributions from silicon photonics, hard disk drive, and packaged-part AI, but minimal silicon carbide or GaN.

• Question from Maxwell Michaelis (Lake Street Capital): And then lastly, I haven't had time to run through the entire press release, but that $5.5 million order you noted in your prepared remarks. Can you go any -- can you share some more detail on that? Is there anything new that we should be looking for? Or is just kind of standard?
  Response: The order includes a mix of AI-related burn-in modules, a large order for new high-power Sonoma systems from a premier test services company, and additional systems for existing customers, with strategic significance for high-volume testing.

• Question from Larry Chlebina (Chlebina Capital): We try to line up your ramp or at least your demand for the systems that you're working on developing for these customers on the AI processors with what's publicly disclosed in terms of the product launch. Is there a case where they may start up on packaged part wherever they have the capacity to do that. And then when they feel comfortable, maybe if it's after the products launched, would they cut over the wafer level burn-in because it's so much more efficient and saves them money? Would they do that? Or would they just do it initially on a brand-new product launch at the beginning?
  Response: It varies by customer and device; some may start with packaged-part burn-in and later transition to wafer-level for efficiency and yield savings, especially for multi-die modules, but wafer-level is increasingly seen as the optimal solution.

• Question from Larry Chlebina (Chlebina Capital): I would think they would need more capacity for the enterprise flash part of it before they ever start needing something for HBF. So the enterprise flash, I'm wondering when is something going to happen there? It seems like it's overdue.
  Response: The customer shifted focus from enterprise flash to the newer, higher-power High Bandwidth Flash (HBF) benchmark, which required significant redesign and slowed progress; the company is ready for production and expects the customer to need more capacity given market trends.

Contradiction Point 1

AI Market Opportunity Characterization

This is a substantial contradiction regarding the **characterization and relative scale** of the AI test market opportunity. In 2025Q4, the CEO made a direct, relative comparison stating the AI market is 3-5x larger than silicon carbide, a key reference point for investors. In 2026 Q2, while providing a specific quantitative estimate ($8-15B), this figure is presented **without reiterating or confirming** the previously emphasized superior magnitude compared to other markets, potentially downplaying its exceptional scale.

Can you provide a mathematical estimate of the long-term AI market opportunity compared to the silicon carbide opportunity? - Christian David Schwab (Craig-Hallum Capital Group LLC)

2025Q4: While AI wafer starts may be about half of silicon carbide's projected 4 million in 2030, the AI market is 3 to 5 times larger due to factors like multiple touchdowns per wafer and varying burn-in times (1 hour to 24+ hours). The market dynamics and potential are significantly larger than the silicon carbide case. - Gayn Erickson(CEO)

Given the significant bookings from the AI processor market, can you provide an indication of whether this business will expand significantly after initial orders? - Christian Schwab (Craig-Hallum)

20260109-2026 Q2: The total AI test and burn-in market is estimated at $8-15 billion. Aehr Test's ASPs are higher on the wafer-level side, and the yield savings advantage is substantial. - Gayn Erickson(CEO)

Contradiction Point 2

Status and Timeline of Key AI Customer Evaluations

This is a substantial contradiction concerning the **status and expected timeline** of a critical customer evaluation process. In 2025Q4, the narrative shifted from evaluation to **active orders and system shipments**, driven by proven feasibility. In 2026 Q2, the same customer's benchmarking is described as **delayed, still pending**, with a new, extended timeline ("next couple of months"), indicating a more protracted and uncertain sales cycle than previously suggested.

What caused the significant shift in tone and excitement in the AI processor market? - Jonathan Dorsheimer (William Blair & Company L.L.C.)

2025Q4: The shift is due to successfully proving the feasibility of wafer-level burn-in for AI processors, moving from evaluation to customer orders and system shipments. - Gayn Erickson(CEO)

Is the delay in benchmark timing at the wafer level due to new product complexities and customers modifying parameters? - Jonathan Dorsheimer (William Blair)

20260109-2026 Q2: The benchmark delay was due to... The data collection is now expected within the next couple of months. - Gayn Erickson(CEO)

Contradiction Point 3

AI Business Growth Timeline and Order Expectations

This is a substantial contradiction regarding the **expected timing for material orders and production ramp**. In 2026 Q1, the company was cautious, stating the timing was **not being specified to avoid guidance** and that evaluations were for processors expected to go into volume production in the second half of "next year" (2027). In 2026 Q2, the tone shifts to assertive visibility, claiming there is visibility on specific GPU, CPU, or ASIC orders, suggesting a near-term order pipeline.

Previously, you were enthusiastic about the TAM for AI-driven products being 3–5x larger than silicon carbide. Is there a timeframe when this will become evident? - Christian Schwab (Craig-Hallum Capital Group LLC)

2026Q1: The timing is not being specified to avoid guidance, but there are evaluations in progress for processors expected to go into volume production in the second half of next year. - Gayn Erickson(CEO)

Given the material bookings from the AI processor market, can you provide any indication of whether the business will expand meaningfully beyond initial orders? - Christian Schwab (Craig-Hallum)

20260109-2026 Q2: The AI business has significant growth potential, with visibility on specific GPU, CPU, or ASIC orders. - Gayn Erickson(CEO)

Contradiction Point 4

Wafer-Level Burn-In Adoption and Customer Transition

This is a substantive contradiction regarding the **technical path and necessity of new products** for a key market transition. In 2026 Q1, the company stated that no new product is needed for qualification, with potential customization. In 2026 Q2, the CEO describes being engaged with customers on "production burn-in at wafer level" as a distinct, separate model from package-level testing, implying it requires a new product cycle and validation to address different parameters.

How should we approach customer engagement during the qualification cycle? - Mark Shooter (William Blair & Company L.L.C., for Jed Dorsheimer)

2026Q1: No new product is needed for qualification; custom WaferPaks may be required based on device pitch. - Gayn Erickson(CEO)

Do you expect ASICs to operate at package level and AI processors at wafer level, or both at wafer level? - Jonathan Dorsheimer (William Blair)

20260109-2026 Q2: The company is engaged with customers on all three models: ... and production burn-in at wafer level. - Gayn Erickson(CEO)

Contradiction Point 5

HBF (High-Bandwidth Flash) Opportunity Scope and Engagement

This contradiction involves the **scale and nature of a significant new market opportunity**. In 2026 Q2, the CEO characterizes the customer's shift to HBF as a "significant new direction that may be larger than the previous enterprise flash application." This suggests a major, potentially **larger market** than the one initially discussed. In the prior quarter (2026 Q1), the HBF engagement was framed as part of the **existing, initial customer discussions** alongside HBM, not as a distinct, larger opportunity.

Does the demand for your AI processor systems align with publicly disclosed product launch timelines, and would customers initially use packaged parts when capacity allows, then switch to wafer-level burn-in post-launch for efficiency and cost savings? - Larry Chlebina (Chlebina Capital)

20260109-2026 Q2: The HBF opportunity is a significant new direction that may be larger than the previous enterprise flash application. - Gayn Erickson(CEO)

You mentioned high-bandwidth flash. Systems makers focus on burn-in due to high motherboard scrap costs. Do you have plans to phase out high-bandwidth memory, or is it mainly high-bandwidth flash? - Bradford Ferguson (Halter Ferguson Financial, Inc.)

2026Q1: The initial engagement was with a lead customer for high-bandwidth flash (HBF), but there are discussions on high-bandwidth memory (HBM) as well. - Gayn Erickson(CEO)