The backlog inflected. The earnings quality did not—yet.
Q4/FY2026 produced record bookings, a threefold FY2027 revenue ambition and a much broader AI test narrative. But the stock now prices in clean execution, while the reported EPS relied on a tax benefit and the operating model is still proving its scalability.
Aehr has moved from a SiC recovery story to an AI-led backlog-conversion story. Record Q4 bookings, $100.6 million of effective backlog and a $130–150 million FY2027 revenue guide are real evidence of demand acceleration. S1
The counterweight is quality and price: Q4’s $0.11 non-GAAP EPS was helped by a $1.846 million tax benefit, customer and delivery concentration are high, and AEHR trades near 17.1× enterprise value / guide-midpoint sales. The burden of proof has shifted from demand discovery to shipment timing, acceptance and margin conversion.
Quality of print: the headline beat outran the operating result
Q4 revenue of $18.835 million beat the four-estimate mean by only $0.150 million, or 0.8%. Standardized adjusted EPS of $0.11 beat the near-zero mean by $0.118 per share—but that EPS comparison is less informative than the income bridge. S1S6
42.6% margin
–6.4% margin
$0.04/share
What is recurring
- Q4 non-GAAP gross margin reached 44.7%, up from 34.4% in Q3, as revenue and mix improved. S1
- Non-GAAP operating income was $0.955 million, a 5.1% margin; non-GAAP pre-tax income was $1.706 million. S1
- Backlog expansion and product mix—not a broad revenue beat—are the durable positives.
- Derived Q4 free cash flow was about $1.7 million, but the full-year figure was negative $5.4 million. S1
What is not yet clean
- The $1.846 million tax benefit equaled 133% of GAAP net income and about 52% of non-GAAP net income.
- The tax benefit contributed roughly $0.056 per diluted share; non-GAAP pre-tax income equaled only about $0.051 per share before a normalized tax rate.
- For FY2026, non-GAAP pre-tax loss was $3.700 million, while a $4.610 million tax benefit produced $0.910 million of non-GAAP net income. S1
- Cash rose mostly because Aehr raised $97.4 million net through equity offerings, not because the business generated cash. S1S5
The guide is backlog-backed, but the conversion path is concentrated
Demand evidence
Q4 bookings were $60.7 million, following $37.2 million in Q3. Effective backlog almost doubled sequentially to $100.6 million; reported backlog rose 108% to $80.6 million. S1S2
Coverage math
Effective backlog covers 67%–77% of the $130–150 million revenue guide and 72% of its midpoint. That is unusually strong visibility, but it is not the same as recognized revenue: delivery, customer acceptance and cancellation terms still matter.
Cadence
Management described a solid Q1, a very strong Q2, roughly flat Q3 and higher Q4; first-half revenue could equal or exceed second-half revenue. Roughly $50 million of Sonoma package-level revenue and 60–70 systems are concentrated around Q2. S3
| Metric | FY26 Q1 | FY26 Q2 | FY26 Q3 | FY26 Q4 | Read-through |
|---|---|---|---|---|---|
| Revenue ($m) | 11.0 | 9.9 | 10.3 | 18.8 | Q4 inflection, not yet a full multi-quarter ramp |
| Bookings ($m) | 11.4 | 6.2 | 37.2 | 60.7 | Two-quarter step-up; H2 bookings exceeded prior guidance |
| Reported backlog ($m) | 15.5 | 11.8 | 38.7 | 80.6 | More than doubled sequentially in Q4 |
| Effective backlog ($m) | 17.5 | 18.3 | 50.9 | 100.6 | Includes orders after period-end; best guide-coverage measure |
| Book-to-bill | 1.04× | 0.63× | 3.61× | 3.22× | Demand intake remains well ahead of shipments |
Sources: company quarterly releases and SEC exhibits. S1S2
Operating trend: revenue and gross margin inflected; GAAP profitability still lagged
Fallback: revenue rose from $10.3m in Q3 to $18.8m in Q4; GAAP gross margin rose from 32.7% to 42.6%; GAAP operating margin improved from –41.0% to –6.4%.
Fallback: adjusted EPS actuals were –$0.01, $0.01, –$0.04, –$0.05 and $0.11. The latest consensus was roughly –$0.01.
GAAP P&L values use company SEC exhibits. The standardized EPS comparison uses Fiscal.ai’s post-print estimate history; its actual series matches company non-GAAP EPS, but the vendor does not explicitly label every historical EPS basis. S1S2S6
The stock celebrated the guide, then surrendered most of the peak
Fallback: AEHR closed at $72.01 before the print, $87.79 the next day, and traded at $77.17 at Jul 20, 2026, 2:09 PM ET. The first-day peak was $110.20.
Reaction scorecard
| Marker | Price | Vs pre-print |
|---|---|---|
| July 14 close | $72.01 | — |
| July 15 open | $98.16 | +36.3% |
| July 15 intraday high | $110.20 | +53.0% |
| July 15 close | $87.79 | +21.9% |
| July 20 intraday snapshot | $77.17 | +7.2% |
The market initially repriced the record backlog and FY2027 scale, with first-day volume roughly 4.9× the 20-day average. By the intraday snapshot, the stock had given back most of the peak response. A sizable short base—4.16 million shares at June 30, roughly 14% of float by secondary-source denominators—likely amplified the first move, but does not explain the fundamental re-rating by itself. S7S8
Fallback: FY2027 revenue consensus rose from $86.85m to $136.27m; adjusted EPS rose from $0.15 to $0.75.
Estimate reset versus official range
- FY2027 revenue mean: $136.3 million, inside the official $130–150 million range and 2.7% below midpoint. S6
- FY2027 adjusted EPS mean: $0.75, based on three contributors; basis comparability with the company’s percentage guide is not fully reconciled.
- FY2028 mean: $201.8 million revenue and $1.38 adjusted EPS, already assuming another year of strong scaling. S6
- Published post-print targets: $110–$125, $115 mean, with four Buy ratings shown by the vendor. Individual estimate timestamps were not supplied in the local feed. S6
Eight Q&A exchanges that change the model
The official webcast archive was cross-checked against a Fiscal.ai/Quartr transcript. The transcript contains obvious “billion”/“million” errors in several guidance passages; official company figures control. S3
Management expects roughly 70% of revenue from AI and 15%–20% from silicon photonics, with the balance from power and other markets. The range is mostly current customers; it includes little or nothing from the new top-tier AI benchmark customer and no memory revenue.
Investor read: the guide is less speculative than the narrative—but benchmark conversion is upside, not base case.
Aehr is engaging two to three flash and two DRAM players, but development requires customer commitment, a purchase order or sponsorship. Management expects no meaningful FY2027 memory revenue.
Investor read: memory is an option on FY2028+, not support for today’s guide.
About $50 million of package-level Sonoma revenue and 60–70 systems are expected, concentrated around Q2. WaferPaks and burn-in modules should represent roughly 30% of total revenue.
Investor read: strong product pull, but one acceptance or component delay can distort a quarter.
The first two devices are package-level opportunities; the second device was delayed but should ramp in FY2027. A third device could move to wafer-level burn-in beyond FY2027.
Investor read: the funnel is deepening, but architecture and schedule remain customer-controlled.
More than 12 GaN WaferPak designs are sampling, while two photonics customers are moving toward production cells and follow-on systems. GaN remains a smaller addressable opportunity than AI or memory.
Investor read: photonics has production evidence; GaN still needs production-order proof.
Management described Q1 as solid, Q2 as very strong, Q3 as roughly flat, and Q4 as higher. The first half could equal or exceed the second half.
Investor read: Q2 is the critical de-risking quarter, not just a seasonal datapoint.
Benchmark performance exceeded expectations. The next step is pilot validation in Taiwan for a current high-volume device plus evaluation of a second device; a FY2027 order could lift revenue above guidance.
Investor read: watch for a purchase order, not another qualification headline.
External capacity can support about 20 Sonoma systems per month and Fremont has substantial theoretical FOX capacity. Actual throughput is constrained by staffing, long-lead parts and power supplies; one key supplier raised quotes about 40% amid NVIDIA-linked demand.
Investor read: “not capacity limited” does not mean execution is frictionless; gross-margin sensitivity matters.
All Q&A items: prepared remarks and analyst Q&A, approximately 29:08–59:39. S3
Industry and positioning read-throughs
Advanced packaging raises the value of early screening
As accelerator packages become more expensive and complex, detecting failures before final assembly improves the economics of wafer-level burn-in. This supports Aehr’s core thesis without identifying any unnamed customer.
Optical interconnect is becoming a second AI leg
Follow-on FOX-XP orders and a newer customer’s multi-wafer cells show photonics moving from evaluation into production. Broader 400G/lane and 1.6T industry roadmaps support the demand backdrop. S10S11
Hyperscalers validate demand but do not guarantee supplier standardization
AI customers are requiring more rigorous screening, yet management said hyperscalers are not broadly directing semiconductor vendors to Aehr. Automotive OEM pull-through appears stronger today.
The same AI boom pressures inputs
Demand from large compute programs validates the test market, but it also competes for Aehr’s power-supply inputs. A 40% quoted increase is a concrete margin and schedule watchpoint. S3
Valuation: the current price already capitalizes a clean ramp
At $77.17, 32.480 million shares imply a $2.506 billion market capitalization. Adding $9.882 million of lease liabilities and subtracting $116.358 million of cash yields about $2.400 billion of enterprise value. Primary statements show no funded-debt line; secondary feeds that label $9.9 million as “debt” are capturing leases. S1S7
| Scenario | Revenue | Illustrative NG margin | Implied NG income | EV / sales | Price / implied income | What must be true |
|---|---|---|---|---|---|---|
| Low guide | $130m | 18% | $23.4m | 18.5× | 107.1× | Backlog converts with some timing/mix friction |
| Mid guide | $140m | 20% | $28.0m | 17.1× | 89.5× | Sonoma wave lands on time and margins scale |
| High guide | $150m | 22% | $33.0m | 16.0× | 76.0× | Current customers execute and upside orders arrive |
| FY2027 consensus | $136.3m | — | $0.75 EPS | 17.6× | 102.9× P/E | Street numbers largely mirror company range |
| FY2028 consensus | $201.8m | — | $1.38 EPS | 11.9× | 55.8× P/E | Another year of scale after the FY2027 step-up |
Low/mid/high margin pairings are illustrative endpoints, not company-provided scenarios. The official release/call basis conflict remains unresolved. Price-sensitive outputs use the intraday snapshot above. S1S6S7
What the stock appears to discount
Near 17× guide-midpoint sales and more than 100× FY2027 consensus EPS, the stock already assumes successful backlog conversion, major operating leverage and continued FY2028 growth. Merely hitting FY2027 guidance may validate the business without expanding the multiple.
What can still create upside
A production order from the benchmark customer, a wafer-level win on the third hyperscaler device, durable photonics follow-ons or a sponsored memory program could push revenue above the base range and extend the runway. Each requires a purchase order or production milestone—not just evaluation progress.
Proof points, break points and the next decision window
What proves the thesis
- Q1/Q2 conversion: Sonoma and FOX deliveries land within the described cadence, without acceptance slippage.
- Recurring margin: non-GAAP operating and pre-tax margin scale toward the guide, independent of tax benefits.
- Customer expansion: the newly benchmarked AI supplier places a pilot or production order.
- Diversification: photonics follow-on systems and SiC orders broaden backlog beyond the two AI customers.
- Cash discipline: working-capital needs rise more slowly than revenue and dilution stabilizes.
What breaks the thesis
- Timing: the Q2 system wave moves out, exposing the concentration embedded in the guide.
- Economics: power-supply inflation, hiring or expedited components prevent gross-margin scale.
- Demand quality: bookings normalize before backlog converts or the benchmark customer remains pilot-only.
- Concentration: any of the three >10% Q4 customers slows; two were AI and one optical. S3
- Capital quality: further equity issuance offsets the per-share benefit of earnings growth.
Catalyst timeline
Confirm unaudited Q4 balance sheet, backlog terms, customer concentration, revenue recognition, tax treatment and final dilution.
Street bar is roughly $26.6 million revenue and $0.09 adjusted EPS; company has not confirmed the report date. S6
Test the roughly $50 million Sonoma expectation, 60–70 system plan and power-supply availability.
Potential benchmark-customer pilot/order, second hyperscaler-device ramp and additional photonics systems.
Third hyperscaler device at wafer level and any sponsored flash/DRAM program.
Capital and governance watch
Aehr ended Q4 with $116.4 million cash and no funded debt, providing ample working capital. But that strength came with dilution: shares outstanding rose 4.9% sequentially and FY2026 net offering proceeds were $97.4 million. S1S5
The company exhausted a $40 million ATM and then announced and completed a new $60 million ATM within nine days in April. That speed improved funding capacity ahead of the ramp, but per-share outcomes should be monitored separately from headline net income.
Evidence ledger and methodology
Primary company and SEC sources control period definitions and reported figures. Third-party feeds are used for consensus, price history, short interest and transcript text. All market data are timestamped; figures may move after publication.
Reported July 14, 2026; quarter and year ended May 29, 2026. Controls reported figures and official guide wording.
Controls historical quarterly trend and prior guidance.
Call held July 14, 2026. Transcript downloaded through Fiscal.ai/Quartr and checked against the official archive and release. No official text transcript was found.
April 8–17, 2026. Controls offering framework; quarterly proceeds derived from official cash-flow tables.
Pulled July 20, 2026; latest estimate-series timestamp July 16. Three-estimate May-end FY2027 series used. Individual contributor timestamps unavailable.
Pulled Jul 20, 2026, 2:09 PM ET. Used for price reaction and price-sensitive valuation.
June 30, 2026 settlement. Exact short-share count is preferred; float percentages vary by provider denominator.
Company releases, July 9 and July 14, 2026.
Industry context only. Broadcom is not identified as an Aehr customer in this report.
The FY2026 10-K had not been filed by July 20, 2026; Q4/FY tables remain unaudited.
Method: calculations are derived from cited source values and rounded. Effective-backlog coverage = $100.6m divided by guide endpoints. Enterprise value = market capitalization + operating lease liabilities − cash. Free cash flow = cash from operations − capital expenditures. This is research, not personalized investment advice.