Fundamentally, Aehr’s current revenue growth, margin and cash flow profile is among the weakest of the names we track across the AI sector, yet the company’s recent order momentum and broadening presence within the AI stack, from GPUs to ASICs to silicon photonics, deserves a deeper look.
Notably, Aehr is pointing for bookings to surge in the second half of its fiscal 2026, or November through May, with management unofficially guiding for $60 to $80 million during this period. This would represent roughly 4X growth from the first half’s $17.6 million in bookings, and supports a strong revenue ramp into 2027 with a flurry of recent orders seeing shipment timelines heavily weighted towards FY27.
Below, we take a look at the bookings growth, plus important comments from management regarding the revenue opportunities in AI and potential for ‘hundreds of millions’ in revenue, what capacity ultimately can support, and the risk of cannibalization.
As those who have been with the I/O Fund for a few years know, AEHR require a more active stance. The stock treated us well in 2022 only because we actively managed the stock, and would only enter again with tight stops and price targets.
Brief Background on AEHR, Main Products
The primary market where Aehr saw former success was with silicon carbide (SiC) in electric vehicles (EVs), but as you can imagine, Aehr's stock has struggled over the past two years as EV sales have declined, causing companies like ON Semi (who supplies Tesla) to cut their orders with the company.
However, Aehr’s stock has seen a resurgence as of late based on its burn-in solutions for AI processors. Similar to EVs, by stress testing the chips at elevated temperatures and voltages (‘burn in’) at the wafer-level, Aehr can lower costs from failures happening at the package or system level. Overall, Aehr’s value proposition is to make sure expensive AI chips don’t fail when placed under high stress.
Testing at the wafer level is attractive due to the complexities of AI hardware, as advanced packages combine multiple GPU or ASIC dies (including dual-die setups) with 8- to 12-high HBM stacks—often totaling dozens of memory dies, all assembled via technologies such as CoWoS. Per Aehr’s CEO: “Even a 0.1% increase in yield by shifting the burn-in of devices from the system or heterogeneous package level to wafer level is very significant.”
Aehr’s main products include its FOX platform for single, multi-wafer and die level burn-in and reliability testing for logic, memory, photonic, and power semiconductors, as well as its Sonoma and Tahoe systems for packaged-part burn-in and high-temp operating life testing. Aehr’s Sonoma system targets leading-edge AI processors/GPUs and networking chips up to 2,000 watts, and allows testing of 22 devices per system.
Combined, Aehr is now the only company that can offer wafer-level and package part burn-in for qualification and production of AI processers with direct, side-by-side comparisons on testing costs, output, operational costs, and impact on yields.
Our coverage on AEHR dates back to 2021 in a deep dive plus further reports in 2022 that discuss the importance of silicon carbide and gallium nitride for lowering power loss in EV batteries.dates back to 2021 in a deep dive plus further reports in 2022 that discuss the importance of silicon carbide and gallium nitride for lowering power loss in EV batteries.
For more on Aehr’s background, acquisition of Incal and the Sonoma systems, refer to our September analysis here: Aehr Test Systems: Optimism Driven by Sonoma Follow-on Orders Despite Soft FY25.Aehr Test Systems: Optimism Driven by Sonoma Follow-on Orders Despite Soft FY25.
Aehr Outlines $60-80M Bookings Target for 2H, up ~4X from 1H
Bookings were rather soft in fiscal Q2, down (33%) YoY and (46%) QoQ to $6.2 million, taking 1H bookings to a total of just $17.6 million.
Management explained that while they are not providing formal bookings guidance, they expect bookings in the second half of FY26 (November through May 2026) to be between $60 million to $80 million, based on the strength of recent customer forecasts provided to them. AI is expected to drive a majority of bookings, with packaged-part burn-in the biggest contributor followed by wafer-level, with some contribution from silicon photonics and a tiny portion from SiC.
This would represent roughly 4X of 1H’s bookings at the $70 million midpoint, and 2-3X its 2H revenue guidance of $25-30 million. Aehr had revealed in Q2’s call in early January that during the first six weeks of Q3, they had received an additional $6.5 million in bookings, already surpassing Q2’s total.
This bookings guidance should not be overlooked in the slightest, should it materialize, as it would represent record bookings by nearly 2X on the low end versus Q3 FY23’s record of $33.3 million. Additionally, looking at the lens of bookings to revenue supports management’s view that this guide should translate into a “very strong” FY27 (starting end of May).
Although bookings have been quite lumpy over the last few years, with a handful of large QoQ fluctuations, the broader trends is fairly clear, in that revenue typically lags bookings by one to two quarters, directionally speaking. An important note on the chart below – the near-vertical move reflects 2H’s guidance, including both Q3 and Q4, so growth will be spread across both quarters rather than hitting all at once.
At a quick glance, Aehr’s current FY27 revenue estimate sits at $82.1 million, up 72% YoY, with the company likely exiting FY26 with strong visibility into a majority of next year’s revenue, adding a layer of confidence in next year’s growth panning out.
Strong Order Momentum in 2026
Underpinning this unofficial bookings forecast (and implying it has a strong chance of coming to fruition) is strong order momentum across the AI stack, from leading AI ASICs customers to silicon photonics players and further upstream to testing and assembly firms. Below offers a brief recap of the recent orders that Aehr has announced as well as commentary from Q2’s call on additional other customer opportunities.
- Aehr announced on March 31 that it had received an initial order from a major new silicon photonics customer, spanning FOX-XP wafer-level burn-in, a fully integrated WaferPak Auto Aligner, and multiple sets of FOX WaferPak full-wafer contactors, with shipments in Q4 (likely May based on Q2’s commentary). Follow-on orders could potentially arrive later in calendar 2026 as this customer’s capacity ramps.
- Aehr announced on March 3 that it had received a follow-on order from its lead silicon photonics customer for one new FOX-XP platform and an upgrade of an existing FOX-XP platform to the new fully automated configuration, shipping in the second half of calendar 2026.
- Aehr announced on February 26 that it had received a $14 million order from its lead AI processor customer for multiple new fully automated FOX-XP systems, a set of FOX WaferPak full-wafer contactors and a fully integrated FOX WaferPak Auto Aligner. Aehr said the systems would ship in the next six months.
- Aehr announced on February 11 that it had received an initial production purchase order from its lead packaged-part burn-in customer for its next-gen AI ASIC chip. This order includes multiple Sonoma systems, fully turnkey burn-in modules (BIMs) and device-specific sockets. This customer is also “forecasting a very large expansion of Sonoma system purchases for that device in the second half of calendar 2026 and continuing into 2027,” per Aehr. To note, this is the customer Aehr was referencing in Q2’s call regarding the ‘substantial forecast’ it had provided for ASIC production capacity with requested Sonoma delivery starting in Q1 FY27 (May-July), supporting its “record bookings” and strong revenue growth outlook for FY27.
- Aehr announced on January 8 that it had received orders from multiple customers for its Sonoma systems, totaling $5.5 million, with management noting that this already is outpacing Q2’s Sonoma volumes.
Other Customer, Product Opportunities
There were a handful of other customer engagements and future opportunities discussed last quarter, though these are either small in nature or with contributions more than a year in the future.
Aehr disclosed that they have two other AI processor firms planning wafer-level benchmark evaluations, which typically take six months, with “meaningful progress” beginning in FQ3.
Management also noted that they proposed a next-gen solution for testing high-bandwidth flash (HBF) leveraging FOX-XP, WaferPaks and auto-aligners, though development of this solution will take more than one year.
Aehr also said that it is installing additional FOX-CP systems at a major HDD supplier for special component burn-in, with additional purchases later in calendar 2026, but added that the “overall revenue opportunity remains modest due to short stress times and the massive parallelism achieved on our FOX-CP system. and proprietary high-power WaferPak wafer contactors.”
For a quick note on SiC, management said demand is weighted toward the end of FY26, and while customers remain optimistic about capacity needs, Aehr is taking a more conservative stance of “show us the orders before we believe them.” Management added that their lead SiC customer is “seeing additional needs for WaferPaks this year, but additional capacity for systems appears to be a year out”
Single AI Processor Could Drive up to $150M Revenue Opportunity
Considering the strength of demand Aehr is seeing recently across the AI ecosystem and strong order momentum for shipments to support AI processors extending through this year into next, Aehr’s commentary on the broader revenue opportunity that AI chips offer is quite important. Management explained this quarter that they are still figuring out just how large this opportunity could be, but revealed that a single AI processor engagement could be worth as much as $150 million.
CEO Gayn Erickson explained that Aehr is “still trying to get our arms around how big it is. What we get is visibility of a specific GPU or CPU or network processor or an ASIC. And then we hear these things from the customer and then we look externally and what are they telling the Street and try and correlate to those lookups. And I'd say pretty consistently, we hear bigger numbers from the customer than the Street. … But a single processor for some of these big guys at wafer-level burn-in is 20, 30 systems or so. And these are $4 million, $5 million machines. So you get a feel for the size of what that looks like.”
Though there is a bit of a range here, Aehr is essentially saying that the largest AI processor companies, such as Nvidia and Broadcom, are driving wafer-level burn-in demand of $80 to $150 million per a single processor. Aehr noted that AI spend in testing between test and burn-in is estimated to be between “$8 billion, $10 billion to maybe $15 billion or so,” meaning these opportunities are roughly just 1% of the broader market size.
Erickson added, “So can the AI business be measured in hundreds of millions of dollars for Aehr Test a few years out? Yes. for sure.” While it is not simply so straightforward to assume that Aehr will ramp to hundreds of millions in revenue over the next few years, it reinforces the company’s value proposition in serving both packaged-part and wafer-level burn-in, improving yield and reducing defects of leading AI chips.
Aehr Hints that Capacity Could Support Much Higher Revenue
While the bookings forecast certainly is a bright spot considering the fundamental picture has been quite challenged, almost more important was management’s commentary on capacity, as it supports the long-term potential to reach hundreds of millions in revenue.
When asked directly about annual manufacturing capacity for wafer-level systems, CEO Gayn Erickson responded that Aehr has discussed with customers “about capacities exceeding 20 systems a month at either package or wafer level. If we had to, we could ship 20 systems a month of each during this calendar year. Now that's bigger than our forecast by a lot. But you know what, when people are saying, could you do something like this and intercept something, it's like if they gave you an order for 50 or 100 Sonomas, like how long is it going to take you to build them?”about capacities exceeding 20 systems a month at either package or wafer level. If we had to, we could ship 20 systems a month of each during this calendar year. Now that's bigger than our forecast by a lot. But you know what, when people are saying, could you do something like this and intercept something, it's like if they gave you an order for 50 or 100 Sonomas, like how long is it going to take you to build them?”
Putting this into context offers a handful of key takeaways, notably that shipping at full capacity each and every month would imply revenue likely 10X higher than what it is today. Aehr had explained that wafer-level burn-in systems such as FOX do carry higher ASPs, and based on the $4-5 million/system comment from above, 20 systems monthly implies maximum annual revenue of $960 million to $1.2 billion. Adding in Aehr’s packaged-part testing systems, which are implied to carry lower ASPs, could still unlock hundreds of millions in annual revenue on top of that.
On the other hand, reading in between the lines of Erickson’s answer for either wafer or package level implies Aehr’s current manufacturing footprint is only supporting 20 systems in total per month, while a step up to 20 of each per month would likely incur higher operating expenses and potential margin impacts. This was hinted at in a later answer, with Erickson saying Aehr is “not happy with these revenue levels, right? We're not making money at these levels. But we would be making more money. We're spending money. We got our foot on the gas. And in fact, it's our expectation that we'll increase the R&D spend particularly in the AI wafer-level burn-in, a little bit in the package because we spent a lot of money on that in just this last year for package, getting this new product out, and then the memory system which will be a blade in our FOX system basically.”
Cannibalization of Wafer-Level Burn-in and Packaged-Part Burn-In
While there was plentiful discussion on some of the challenges that Aehr and its customers are facing when it comes to evaluating wafer-level burn-in systems (such as not being fully aligned on parameters when benchmarking, thus causing some delays), the main takeaway on the product side from last quarter’s Q&A was that there will be cannibalization between FOX and Sonoma.
William Blair’s Jonathan Dorsheimer questioned about the potential for cannibalization, noting that the AI processor customer is moving forward with wafer-level testing, yet the ASIC customer is on packaged-part, and if future deployments would remain like this or both shift to wafer-level.
As AI chip design now shifts to more advanced packages combining multiple GPU or ASIC dies (including dual-die setups) with 8- to 12-high HBM stacks, often totaling dozens of memory dies, assembled via technologies such as CoWoS, Erickson pointed out that most qualification is done on the full package. He added that some customers now “would like to be able to qual the processor inside when it's still in wafer form [because] from a production perspective, the value proposition is you're burning in these devices and when they fail, you take out the other compute chip and all the memory plus the CoWos substrate” to avoid scrapping expensive, supply-constrained components during the development process.
However, the key takeaways is that if customers “could move it to wafer level, [they] don't need to do it a package anymore,” which Erickson says will “for sure” mean there is cannibalization between FOX and Sonoma. However, he believes that the “the world is going to be both for a long time, and we're in a great position to do both.” While it remains to be seen if/when cannibalization occurs at its lead customers, it’s a risk to be aware of.
Financials
Revenue Growth to Accelerate from FQ4
Aehr’s FQ2 revenue ending November was down by (26.5%) YoY and (9.9%) QoQ to $9.9 million. The decline was primarily driven by lower WaferPak shipments, partially offset by stronger demand for the Sonoma systems. Management pointed out that they experienced some GaN semiconductor headwinds, delaying ~$2 million in WaferPak shipments from Q2 to Q3.
Systems revenue grew by 46% YoY and down (26%) QoQ to $5.0 million. Contactor revenues, which include WaferPak contactors, were down (60%) YoY and up 32% QoQ to $3.44 million due to the ongoing softness in demand for electric vehicles and partially offset by the increase in package-level burn-in systems revenue, driven by increased demand from customers in AI-related applications.
Aehr will report its Q3 FY2026 ending February results on April 07th. Revenue is expected to be down (40.8%) YoY to $10.9 million. This is the last quarter of negative YoY growth as the company has diversified from silicon carbide for electric vehicles to fast growing AI and data center infrastructure markets.
Management highlighted during the earnings call that this year they are making significant progress expanding into additional key markets for the semiconductor test and burn-in solutions, including AI processors, gallium nitride power semiconductors, data storage devices, silicon photonics integrated circuits and flash memory, which will help the company to grow its revenues and profits.
Analysts expect strong revenue growth in the coming quarters. Revenue is expected to accelerate to 14% YoY in FQ4, then to 43.9% and 88.6% in the next two quarters.
Looking ahead, due to the strong AI-related demand, the company’s FY2027 revenue ending May is expected to ramp significantly and accelerate to 72% YoY to $82.1 million from an expected decline of (19%) for FY2026.
Margins
The company’s margins are under pressure due to the lower revenue and unfavourable product mix.
- FQ2 gross profits were $2.6 million or 25.7% of revenue compared to $5.4 million or 40.1% in the same period last year. The adjusted gross margin was 29.8% compared to 45.3% a year ago. The decline was due to lower revenue and a less favorable product mix as last year's quarter included a higher proportion of higher-margin WaferPak revenue.
- FQ2 adjusted operating loss was ($2.7 million) or (27.6%) of revenue compared to adjusted operating income of $0.2 million or 1.5% of revenue in the same period last year. The decrease in margins was primarily due to lower revenue.
- FQ2 adjusted net loss was ($1.3 million) or (13.1%) of revenue compared to adjusted net income of $0.70 million or 5.1% of revenue in the same period last year.
Positive Adjusted EPS Expected in FY2027
Analysts expect adjusted EPS of ($0.07) for Q3 FY2026 and expect to improve in the coming quarters. Adjusted EPS is expected to be ($0.13) for the FY2026 ending May and to improve significantly to $0.13 in FY2027 and $0.26 in FY2028.
Cash Flow and Balance Sheet
The company’s cash flows have been weak due to current losses. However, it should improve in the coming quarters as the revenue is expected to ramp significantly in FY2027.
- FQ2 operating cash outflow was ($1.17 million) or (11.9%) of revenue compared to ($5.9 million) or (43.7%) of revenue in the same period last year. Cash outflows were higher last year due to higher supplier payments.
- FQ2 free cash outflow was ($1.64) million or (16.6%) of revenue compared to ($6.2 million) or (46%) of revenue in the same period last year.
- The company had cash of $30.8 million and no debt at the end of FQ2 compared to $22.7 million and no debt at the end of the previous quarter. Cash increased as the company raised $10 million in gross proceeds through the sale of about 384,000 shares.
- Inventory increased slightly by 2.2% QoQ to $42.7 million.
Conclusion
Aehr’s fundamentals are challenged, no doubt, but the company’s accelerating order momentum and expectations for bookings to surge through the end of May support a strong revenue ramp into 2027, currently estimated at 72% YoY to $82.1 million. Aehr’s value proposition of offering both wafer-level and packaged-part burn-in testing across the hardware stack, helping improve yield and ensuring highly-complex, expensive AI chips do not fail under high stress, is a key driver of this strong product demand.
Given the weak fundamentals at the moment, we will treat Aehr as a momentum trade for the time being as we await further confirmation with material evidence in earnings reports that bookings and revenue growth are materializing as expected.
Please note: The I/O Fund conducts research and draws conclusions for the Fund’s positions. We then share that information with our readers. This is not a guarantee of a stock’s performance. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis.
Damien Robbins and Royston Roche, Equity Analysts at the I/O Fund, contributed to this analysis.
Recommended Reading: