MRAM Demand Is Delivering, but the Cash Payoff Is Further Away
Key Takeaways
- Product sales drove the step up in revenue. MRAM product sales rose 33.5% year over year and supplied most of the sequential increase. The $22M Q4 guidance midpoint is $1M above our prior estimate.
- Product gross margin fell even as sales grew. Product gross margin fell to 43.2% from 45.8%, and the defense subcontract helped keep consolidated margin above 50%. Q4 price increases now have to show up in gross profit.
- The cost of getting to cash profitability rose. Litigation reached $5.987M, above the roughly $4M expected, and a separate $2.715M export-control provision contributed to a $0.29 GAAP loss per share.
- Rating: Maintaining Outperform. We cut 12-month fair value to $17–20 from $18–22 and lower conviction to 5/10. At $14.43, stronger shipments and signed defense work still offer enough upside to compensate for weaker margins and cash conversion.
Results vs. Consensus
Everspin’s preliminary results cover the quarter ended September 30. Q2’s case rested on a product-led step up with broadly stable nonproduct revenue. That is what the sales mix delivered. The same quarter showed why stronger demand alone cannot restore the prior valuation: manufacturing costs absorbed much of the product growth, and legal costs rose faster than management had expected.
| Q3 FY2026 metric | Actual | Street consensus | Company guidance | Result |
|---|---|---|---|---|
| Revenue | $20.648M | $20.07–20.10M | $19.5–20.5M | Beat by 2.7–2.9%; $0.148M above guide |
| GAAP diluted EPS | $(0.29) | n/a | $(0.10) to $(0.05) | $0.19 below the low end |
| Non-GAAP diluted EPS | $0.12 | n/a | $0.10–0.15 | Within guide |
| GAAP gross margin | 51.4% | n/a | n/a | Down 2.5 pp sequentially |
| GAAP operating loss | $(7.929)M | n/a | n/a | Loss widened sequentially |
| Operating cash flow | $0.863M | n/a | n/a | Up from $0.159M in Q2 |
| Free cash flow after PP&E and intangible purchases | $(1.043)M | n/a | n/a | Still negative |
Year-over-Year Comparison
| Metric | Q3 FY2026 | Q3 FY2025 | Change |
|---|---|---|---|
| Revenue | $20.648M | $14.060M | +46.9% |
| GAAP gross profit | $10.609M | $7.206M | +47.2% |
| GAAP gross margin | 51.4% | 51.3% | +0.1 pp |
| GAAP operating expenses | $18.538M | $8.751M | +111.8% |
| GAAP operating loss | $(7.929)M | $(1.545)M | Loss widened $6.384M |
| GAAP net income (loss) | $(7.175)M | $0.054M | Down $7.229M |
| GAAP diluted EPS | $(0.29) | $0.00 | Down $0.29 |
| Non-GAAP net income | $3.058M | $1.461M | +109.3% |
| Non-GAAP diluted EPS | $0.12 | $0.06 | Up $0.06 |
Quarter-over-Quarter Comparison
| Metric | Q3 FY2026 | Q2 FY2026 | Change |
|---|---|---|---|
| Revenue | $20.648M | $18.736M | +10.2% |
| GAAP gross profit | $10.609M | $10.097M | +5.1% |
| GAAP gross margin | 51.4% | 53.9% | −2.5 pp |
| GAAP operating expenses | $18.538M | $14.480M | +28.0% |
| GAAP operating loss | $(7.929)M | $(4.383)M | Loss widened $3.546M |
| GAAP net loss | $(7.175)M | $(3.589)M | Loss widened $3.586M |
| GAAP diluted EPS | $(0.29) | $(0.15) | Down $0.14 |
| Non-GAAP net income | $3.058M | $2.862M | +6.8% |
| Non-GAAP diluted EPS | $0.12 | $0.11 | Up $0.01 |
| Cash and equivalents | $43.042M | $43.896M | Down $0.854M |
Revenue Assessment
The service ramp drove most of Q2’s sequential increase. In Q3, the larger contribution came from products as energy designs moved to shipments alongside stronger demand from established industrial and data center customers. The stronger evidence for our FY27 estimate is that products grew alongside defense services, rather than needing another step up in contract recognition.
Margin Assessment
The stable year-over-year consolidated margin masks a roughly 4.6 percentage point decline in product margin, offset by a greater share of high-margin nonproduct revenue. The sequential comparison is a cleaner warning: nonproduct revenue stayed near 18% of sales and company margin still fell 2.5 points. We reduce our FY27 product margin assumption to 46% from 48%; the forecast gives Q4 pricing some credit while allowing for continued higher input costs.
Earnings and Cash Assessment
| Q3 earnings reconciliation | $M |
|---|---|
| GAAP net loss | (7.175) |
| Stock compensation added back | 1.531 |
| Litigation added back | 5.987 |
| Regulatory provision added back | 2.715 |
| Non-GAAP net income | 3.058 |
Adjusted operating income increased to $2.304M from $2.068M in Q2. That modest increase on nearly $2M of extra sales is a better measure of incremental operating leverage than the doubling of adjusted EPS from a year ago. The latter comparison includes a prior-year adjusted operating loss and masks the two current legal items.
Q3 operating cash flow improved to $0.863M, but $1.906M of PP&E and intangible purchases kept free cash flow negative. Options and employee share purchases provided $0.205M and finance leases used $0.016M, leaving cash down $0.854M to $43.042M. The nine-month picture is more revealing: $1.592M of operating cash fell well short of $8.559M of investment, even with support from payables and accrued liabilities. Those balances cannot be relied on as permanent funding. We value the cash Everspin is likely to retain after investment and settlement costs.
Segment Performance
Everspin reports product and nonproduct revenue categories. Its operating discussion identifies demand by end market without disclosing separate end-market sales or profit.
| Revenue category | Q3 FY2026 revenue | YoY | QoQ | Gross margin | Q2 margin | Q3 FY2025 margin |
|---|---|---|---|---|---|---|
| MRAM products | $16.945M | +33.5% | +10.7% | 43.2% | 45.8% | 47.8% |
| Licensing, royalties, engineering services and other | $3.703M | +170.3% | +8.2% | 88.7% | 90.0% | 83.5% |
| Total | $20.648M | +46.9% | +10.2% | 51.4% | 53.9% | 51.3% |
MRAM Products: Adoption Extends Beyond the Industrial Recovery
“Growth in Industrial Automation was led by a continued recovery in customer demand, particularly in Japan. In the Data Center market, we saw strong demand for RAID storage configurations, while Energy Management benefited from the transition of new design wins to shipments in Europe.” — Sanjeev Aggarwal, CEO
Industrial recovery and repeat data center demand have a different investment implication from new European energy shipments: the first two can lift near-term volumes without adding a new socket; the last converts a design win into production. Medical demand offers another route for adoption. In infusion pumps and dialysis equipment, memory must retain calibration, dosing or sensor data through power interruptions and repeated writes, which gives MRAM a use even where it cannot compete with flash on cost alone.
Assessment: We raise FY27 product sales to $74M from $72M. The increase reflects broader shipments of existing products, with room for a slower industrial recovery; it assumes neither a new CXL platform nor high-density UNISYST volume.
Licensing and Services: The Contract Helps More at Gross Profit Than at Revenue
“Licensing, royalty, engineering services and other revenue increased to $3.7 million from $1.4 million in Q3 '25, primarily due to continued revenue recognition on our $40 million subcontract agreement with a U.S. prime contractor.” — Bill Cooper, CFO
The category earned $3.286M of gross profit, or about 31% of consolidated gross profit on 18% of sales. The existing contract is doing the work the Q2 thesis assigned it. The older $14.6M sustainment program is separate and recorded in other income. Its continued wind-down tempers the net earnings benefit of the new work, even as the new contract lifts operating earnings.
Assessment: We keep FY27 nonproduct revenue at $14M. The quarter supports a meaningful contribution from signed work, but a blended category margin near 90% is too rich to carry through every engineering milestone.
Key Topics & Management Commentary
Overall Management Tone: Management sounded more assured than in Q2 about converting designs into shipments and more willing to act on pricing. Its confidence was narrower on the earnings payoff: it gave a direction for legal spending and margins, while leaving the size and durability of the improvement open.
1. Pricing Becomes the Test of Product Economics
Q2’s weaker product economics looked like an input-cost problem that could persist. Q3 confirmed it: product revenue rose, yet manufacturing, materials and services costs reduced gross profit per sales dollar.
“To counter rising input costs, we began to institute pricing increases in the fourth quarter and expect product margins to improve sequentially.” — Bill Cooper, CFO
Assessment: We expect Q4 product margin to recover to 45%, still below Q2. The burden is on realized pricing: a failure to improve from 43.2% despite higher forecast sales would lower our FY27 estimate and make the richer end of the valuation range hard to defend.
2. Litigation Exceeds the Prior Allowance
Patent litigation reached $5.987M after $4.027M in Q2, breaching the expense assumption we set last quarter. The separate export-control reserve is a potential settlement payment; it does not tell us when the quarterly cost of the patent case will fall.
Assessment: We expect $5.1M of Q4 litigation expense and $7M in FY27. That is a meaningful reduction from FY26 but still leaves FY27 GAAP earnings near break-even. A procedural milestone can reduce spending without ending the case. The export-control matter has a different path and remains a separate cash risk.
3. Teledyne Widens the Aerospace Route to Market
“Through the partnership, Teledyne HiRel will offer our 64-megabit, 128-megabit and 256-megabit PERSYST STT-MRAM solutions as part of its memory products portfolio for military and aerospace customers with target applications such as avionics, VPX and single-board computer platforms, radar and electronic warfare payloads, satellite electronics and autonomous systems.” — Sanjeev Aggarwal, CEO
Assessment: The partnership makes the PERSYST density expansion announced in Q2 more useful: a specialist channel can put the parts in front of customers with long qualification cycles. Everspin also reports space designs progressing into production, but no separate STT revenue or order value. We credit the route to market in the growth outlook and leave a large new aerospace program out of the forecast.
4. NOR Scarcity Opens a Door Before UNISYST Is Ready
Distributors are already asking for existing MRAM parts to replace scarce NOR flash. That is a shorter path to revenue than waiting for the new high-density family. The two opportunities solve different problems: availability can win a socket now; repeat volume after NOR supply recovers requires a strong cost or performance case.
“We remain on track to have a 256-megabit xSPI prototype based on TSMC's 16-nanometer FinFET process technology in the first half of 2027. This prototype will verify our design and put us on the path to designing high-density UNISYST, unified code and data memory products, later in the year and beyond.” — Sanjeev Aggarwal, CEO
Assessment: Q&A pointed to Q1 for the prototype, within the broader H1 window in the prepared remarks. That is broadly consistent with the revised early-2027 engineering milestone but leaves the earlier late-2026 tapeout and a material production ramp unconfirmed. We allow some benefit from current parts in the Q4 midpoint and still assign no material FY27 revenue to UNISYST.
5. Microchip Qualification Now Has Two Milestones
“We remain on schedule with our goal of delivering first qualified silicon for our Toggle and sensor products by the second half of 2027 and STT products by the first half of 2028.” — Sanjeev Aggarwal, CEO
Assessment: Process transfer and mask work give the second fab a clearer path to qualification. Staggering the milestones makes STT the later test and preserves the gap between qualified silicon and commercial output. We still assume no material FY27 contribution and keep post-2028 Chandler access open: the earlier product family’s milestone alone cannot remove the risk for the full portfolio.
6. CXL Clears the Demo; LPDDR Extends the Pipeline
Everspin met the September CXL demonstration commitment. A persistent buffer on the standard CXL interface could keep small, frequent writes available through an interruption without first flushing to an SSD. The next economic test is whether a customer workload saves enough time or complexity to justify another memory tier.
“So since then, we've actually had several follow-on discussions trying to identify use cases. We obviously don't have anything yet, but we're hoping that we will have something identified here in Q4 or Q1.” — Sanjeev Aggarwal, CEO
Assessment: We give the demo credit for lowering integration risk and look for a customer-selected use case next. The JEDEC effort addresses a different barrier, controller and vendor compatibility for persistent LPDDR. Management’s potential end-2028 sample and early-2029 revenue milestones put that effort outside the forecast. Neither project offsets near-term product cost pressure.
Guidance & Outlook
| Metric | Prior Q3 guide | Q3 actual | New Q4 guide | Q4 Street consensus |
|---|---|---|---|---|
| Revenue | $19.5–20.5M | $20.648M | $21.5–22.5M | $20.87–20.88M |
| GAAP diluted EPS | $(0.10) to $(0.05) | $(0.29) | $(0.13) to $(0.08) | n/a |
| Non-GAAP diluted EPS | $0.10–0.15 | $0.12 | $0.13–0.18 | $0.12 |
The $22M midpoint calls for 6.5% sequential revenue growth and sits about 5.4% above the Street. We expect another product-led quarter: $18.3M of products and $3.7M of nonproduct revenue. The company has already included some NOR replacement demand in its guide. We also expect Q4 price increases to contribute, but see little reason to extrapolate Q2’s outsized beat: Q3 cleared the ceiling by only $0.148M.
The adjusted EPS midpoint rises to $0.155 from the $0.125 midpoint of the old Q3 guide, and sits $0.035 above the Q4 Street estimate. The GAAP range, by contrast, is worse than management’s prior Q3 range, even though it improves on the actual loss. Q4 adjusted earnings exclude litigation, Microchip engineering fees and stock compensation. Everspin can meet its adjusted guide through growth and pricing without doing enough to cover those costs on a GAAP basis.
Analyst Q&A Highlights
What the Year-End Legal Milestone Could Change
The question asked how much litigation expense Q4 absorbs and what that means for next year’s operating expenses.
Q: “My first question, I wanted to ask on litigation costs. So it looks like they stepped up in the quarter. I was hoping maybe you could discuss sort of what the cost assumption is embedded into the 4Q '26 guide. And then maybe, if possible, could you discuss OpEx directionally as a whole as we start to look into next year?”
— Neil Young, Needham & Company
A: “What I would say is in Q4, we expect to see some still fairly substantial costs. I would expect that it is going to be slightly lower than what we saw in Q3. And then as we look forward in 2027, one of the things we did want to stress is this matter will go to the ALJ at the end of this year. And then as we move and look forward into 2027, those litigation expenses should reduce significantly.”
— Bill Cooper, CFO
Assessment: The answer identifies why spending may fall and roughly when; it does not establish total FY27 operating costs. Our adjusted cost base still grows, so the assumed legal relief cannot all flow through to incremental GAAP earnings.
The Export-Control Fine Is Still Open
Q: “You mentioned the regulatory provision. Can you explain what this is and where this comes from?”
— Richard Shannon, Craig-Hallum
A: “So on that one, we actually have a fine that's related to export controls from BIS, a government agency within the Department of Commerce. And so we took that particular provision in anticipation of settling that particular fine agreement. It's not settled yet, but that's our best estimate today of what we expect to close it out.”
— Bill Cooper, CFO
Assessment: The provision estimates the possible payment; it does not establish a settlement amount or date. We allow for cash at the accrued amount and leave any additional terms outside the base case. Unlike the patent expense, the provision opens a separate risk around export compliance.
Where the Product Cost Pressure Sits
Q: “So, Bill, I think you mentioned in regards to the third quarter margins that you had a positive mix offset by some manufacturing costs. I wonder if you could help us understand both of those dynamics here. Is the mix here product or end market or geography related? And then what are these manufacturing costs? And what's the sustainability of that dynamic?”
— Richard Shannon, Craig-Hallum
A: “And we had signaled, right, that we were going to see some -- continue to see some product headwinds in terms of the cost, right, materials and services costs. And obviously, there's a lot of pieces that go into that bucket. And we recognize that, right? We had seen it coming. And for that reason, we instituted a price increase in Q4 as well. We continue to see a good product and nonproduct revenue mix. I think really, we were trying to indicate that, again, those headwinds are really confined to the product side of the equation.”
— Bill Cooper, CFO
Assessment: The answer separates the high-margin revenue mix from the costs inside products. It does not say which products or geographies can pass on more of the increase. The Q4 margin is therefore a better test of pricing power than a recovery in the consolidated number driven by services.
Nonproduct Margin Depends on What Is Delivered
Q: “From a segment perspective, licensing, engineering and other revenues picked up strongly as expected from the recent contract. It was good to see. But what I noticed is that the gross margin in that segment also stayed relatively stable, very strong, close to 90%. Is there any -- is that sort of a reasonable expectation going forward? Or just wondering how much visibility you have given this new revenue stream?”
— Jack Vander Aarde, Maxim Group
A: “So you can kind of think of it in terms of the 70%, 80%, up to 90% range. Certainly, again, it depends on a number of factors depending on whether it's royalty or foundry services or licensing or engineering services. That's really what that mix of that really drives the margin.”
— Bill Cooper, CFO
Assessment: The CFO gave a range for the blended category, rather than a margin promise on the subcontract. We model 85% for FY27, below Q3’s 88.7%. At unchanged annual sales of $14M, a mix shift to 75% would remove $1.4M of gross profit.
How Much NOR Substitution Is in the Guide?
Management said it was hearing of NOR capacity shifting toward DRAM. The follow-up distinguished distributor requests from already material sales.
Q: “Sanjeev, any way that you'd offer to quantify how much of your products that you think could be because distributors are looking to replace NOR Flash they can't find?”
— Richard Shannon, Craig-Hallum
A: “I don't think it's significant or material yet, but we do see traction where it could become significant in Q4 and Q1.”
— Sanjeev Aggarwal, CEO
Q: “Okay. Is that baked into your revenue guide for this fourth quarter?”
— Richard Shannon, Craig-Hallum
A: “It is, yes.”
— Sanjeev Aggarwal, CEO
Q: “Any way you try to quantify that for us?”
— Richard Shannon, Craig-Hallum
A: “We've typically not done that. So let me get back to you on that one.”
— Sanjeev Aggarwal, CEO
Assessment: The company has allowed for substitution demand in its sales forecast but declined to size it. A shortage creates a reason to try the part; it gives us no basis to assume repeat demand at the same price once flash is available again.
Design Wins: Real Shipments, Limited Pipeline Visibility
Q: “For design wins overall just tracking this year, are they up overall? Do you have a good sense of that? And just sort of the size and scope of those opportunities?”
— Jack Vander Aarde, Maxim Group
A: “So, we usually report our annual design wins at the end of the year in our 10-K. But I think a good indication is the product revenue growth that you're seeing quarter-over-quarter and also year-over-year, that basically indicates the number of design wins that we've been able to turn into production and the new design wins that we're seeing.”
— Sanjeev Aggarwal, CEO
Assessment: Sales confirm that some designs converted. They cannot tell us how many new wins replace the ones now shipping or how large those programs are. Our estimate slows product growth next year rather than projecting the current quarterly pace across the pipeline.
LPDDR Standards Set the Earliest Commercial Window
Q: “And then on the JEDEC task group, what's the time line for kind of setting standards or those conversations to maybe translating into revenues for you guys or at least something that's observable?”
— Josh Sullivan, JonesTrading
A: “So we expect that we will probably reach some kind of agreement with all the participants over the next 12 to 15 months, and that will probably take another quarter for that spec to get published. And our goal is to start the design early next year so that we are actually ready to intersect when this spec is actually published. So 6 months later, hopefully, we'll have samples on hand, which would be an LPDDR6 MRAM that we can then start sampling with our customers. So what I would say is samples by the end of '28, revenue sometime -- early revenue in '29.”
— Sanjeev Aggarwal, CEO
Assessment: Designing while the specification is being written could shorten time to market. It also makes the sample schedule contingent on industry agreement. The earliest milestones we can test are a design start and an agreed specification, well ahead of any earnings contribution.
What They’re NOT Saying
- How much pricing sticks. Everspin did not quantify coverage or realized pricing. Margin recovery will depend on the price it collects after customer negotiations and on whether input costs keep rising.
- Contract-level margins and future milestones. The mix range is helpful, but the company did not disclose a margin or recognition schedule for the remaining subcontract work. That keeps quarterly gross profit sensitive to the timing and mix of deliveries.
- Terms of the export-control resolution. The provision is a dollar estimate. The timing, final payment and any non-monetary requirements remain open.
- What qualifies UNISYST for sustained volume. The prototype timing is clearer than the handoff to customer qualification and orders. The company also gave no new rail conversion or post-2028 Chandler update.
- Customer economics for CXL. The demo shows connectivity and persistence; management did not report a customer workload gain, design commitment or production schedule. The opportunity still needs a buyer with a measurable problem.
Market Reaction
- Pre-print setup: MRAM closed October 28 at $14.65, up 57.9% year to date and 39.7% over 12 months, but down 19.2% over 30 days. Its preceding 52-week closing range was $7.50–44.01. The S&P 500 was up 13.1% year to date entering the print.
- After hours: Evening indications showed a decline of roughly 10%.
- October 29 reaction session: MRAM opened at $13.54, down 7.6%, traded between $13.54 and $14.95, and finished at $14.43, down 1.5%.
- Participation and benchmark: Volume was about 1.0M shares versus a 0.8M 30-day average, or 1.3 times normal. The S&P 500 declined 0.3%.
The opening gap and partial recovery fit a report that cut both ways. The higher sales outlook gives buyers a reason to look through some of the initial selloff. The product margin and unplanned costs give investors a reason to require a lower multiple even if demand holds. The tape cannot distinguish those explanations from trading flows, but the close gives us a lower entry price than in Q2.
Street Perspective
Debate: Does the Product Ramp Deserve a Premium?
Bull view: Data center and mission-critical applications support repeat demand, and new energy and medical shipments spread growth beyond any one market. Existing product sales can scale before the next generation arrives.
Bear view: The Japanese recovery and NOR scarcity could flatter demand. Pricing introduced to offset cost inflation could reveal how much current growth depends on what competitors have available.
Our take: The move from design wins to shipments supports higher sales, but Q3 incremental product gross profit was only about 19% of incremental product revenue. A durable premium requires a better contribution from each new sales dollar.
Debate: Is Defense Revenue a Stable Profit Engine?
Bull view: The signed subcontract is now delivering a second quarter of meaningful activity, and domestic manufacturing relationships make the franchise harder to replicate.
Bear view: The older program is winding down, future awards are uncertain, and nonproduct gross profit depends on the mix of what gets recognized each quarter.
Our take: Contract execution earns a premium over an unproven pipeline. It does not support a higher nonproduct revenue estimate or the valuation of a recurring royalty stream.
Debate: Will Adjusted Profit Finally Convert Into Cash?
Bull view: Adjusted operating profit is growing, and Q4 pricing plus lower 2027 litigation expense could let more of it reach shareholders. Longer-dated memory projects could add another growth leg.
Bear view: The legal spending path just moved up, the export-control cash payment is unsettled, and manufacturing and product development still need funding. Adjusted profits may keep rising while cash falls.
Our take: We give more weight to the cash gap today. Better pricing and lower legal costs can improve the return, but we want to see both reflected in operating cash before restoring Q2’s valuation.
Our Estimates and Valuation
| Analyst estimate | Post-Q2 | Post-Q3 | Reason |
|---|---|---|---|
| FY26 revenue | $74.6M | $76.3M | Q3 above our forecast and Q4 midpoint higher |
| FY26 GAAP gross margin | 53.7% | 52.5% | Product costs overwhelm mix improvement |
| FY26 litigation expense | $13.7M | $16.7M | Q3 actual and higher Q4 expense |
| FY26 non-GAAP EPS | $0.47 | $0.48 | Q4 sales and a partial pricing recovery |
| FY26 GAAP EPS | $(0.35) | $(0.58) | Litigation and the separate regulatory provision |
| FY26 ending cash | $38M | $39M | Q3 cash held up better than our prior second-half cash path |
| FY27 revenue | $86M | $88M | Existing product shipments; nonproduct sales steady versus prior estimate |
| FY27 non-GAAP EPS | $0.57 | $0.50 | Lower product and nonproduct margin, more core spending and dilution |
| FY27 GAAP EPS | $0.05 | $(0.08) | Lower adjusted earnings, higher litigation and stock compensation |
| Forward net cash for valuation | $35M | $30M | Further investment and settlement cash allowance |
Earnings Bridge
FY26: Reported nine-month sales of $54.256M plus $22M for Q4 produce $76.256M. Our Q4 product and nonproduct sales assumptions carry gross margins of 45% and 88%, respectively, for $11.491M of GAAP gross profit. Adding $0.150M of cost-of-sales stock compensation, deducting $8.300M of adjusted operating expenses and adding $0.700M of net interest and other income after tax gives about $4.041M of Q4 adjusted net income, or $0.16 on 26M diluted shares. H1 and Q3 adjusted net income sum to $8.553M; the full-year result is about $12.594M, or $0.48. In Q4 we retain $5.1M of litigation, $1.5M of stock compensation and $0.4M of Microchip engineering fees for GAAP. That puts the Q4 GAAP loss near $2.96M, or $0.12 per share. Adding it to the nine-month loss yields about $14.02M of FY26 GAAP loss, or $0.58 on 24M shares.
| FY27 analyst bridge | Assumption or result |
|---|---|
| Revenue: product / nonproduct | $74M / $14M; $88M total |
| GAAP gross margin: product / nonproduct | 46% / 85% |
| GAAP gross profit | $45.94M |
| Cost-of-sales stock compensation added back | $0.65M |
| Adjusted operating expenses | $34.50M |
| Net interest and other income after tax | $1.00M |
| Non-GAAP net income / diluted shares / EPS | $13.09M / 26.25M / $0.50 |
| Legal, stock compensation and engineering expense for GAAP | $7.00M / $6.00M / $2.00M |
| GAAP net loss / loss share count / EPS | $(1.91)M / 24.8M / $(0.08) |
We assume about 14% FY27 product growth, below Q3’s year-over-year pace. Adjusted operating expenses of $34.5M are up from the $33.2M annualized Q4 rate as customer and product work expands. We model meaningful relief in litigation and no repeat regulatory charge. We retain our prior $2M of Microchip engineering fees: there was no Q3 addback, but Q4 guidance again excludes those fees and process transfer continues into 2027. A point of product margin is $0.74M of pretax profit, or about $0.03 per diluted share. If litigation instead runs at $14M in FY27, the additional $7M would widen the GAAP loss and cut cash by up to the same amount before any offset.
From $43.042M in Q3 cash, we allow about $4M of Q4 net use, including capex and working capital. In FY27 we expect about $3.5M of operating cash before the assumed BIS payment, against $10M of capital investment. That operating cash estimate starts with the GAAP loss, adds about $9M of stock compensation and depreciation, and allows about $3.6M of working-capital use. A cash settlement at the accrued amount brings FY27 cash use to about $9M. This leaves roughly $30M of forward net cash, assuming no new financing and no funded debt.
12-Month Fair Value
We lower fair value to $17–20 from $18–22. We still value FY27 sales because near-term GAAP earnings are negative and adjusted EPS omits material costs. We reduce the range to 4.7–5.7 times EV/revenue from 5.0–6.2 times. Higher sales support a premium, but thinner product margins, slower cash conversion and the new compliance exposure make the old upper multiple harder to earn. At the $18.58 base case, the stock would already trade at about 37 times our $0.50 FY27 adjusted EPS. We attach no separate value to CXL, UNISYST, additional defense awards or an early Microchip ramp.
Applying 4.7–5.7 times to $88M, adding $30M of forward net cash and dividing by 26.25M diluted shares produces $16.90–20.25 per share before rounding. The lower multiple and cash more than offset the sales increase. We require evidence of better product margins and lower cash legal spending before valuing the company at the high end.
| 12-month scenario | FY27 sales | EV / sales | Net cash / diluted shares | Value per share | Total return from $14.43 |
|---|---|---|---|---|---|
| Downside: pricing fails, services slow, legal costs persist | $77M | 3.0x | $24M / 26.5M | $9.62 | −33% |
| Base: products scale and margins partly recover | $88M | 5.2x | $30M / 26.25M | $18.58 | +29% |
| Upside: faster product growth and better cash conversion | $96M | 6.2x | $36M / 26.25M | $24.05 | +67% |
We raise the downside weight to 30% from 25%, retain 50% for base and reduce upside to 20%. The result is about $17 of weighted value, or 18% expected total return with no dividend. That clears the same illustrative 8% S&P 500 hurdle we used in Q2, but by less than before. The downside is credible if end-market recovery fades before higher prices take hold and legal bills remain elevated.
Thesis Scorecard Post-Earnings
| Standing thesis point | Status | Quarterly assessment |
|---|---|---|
| Bull 1: MRAM secular adoption in mission-critical applications | ON TRACK | Medical use and production shipments in energy extend the adoption evidence. |
| Bull 2: Design wins translate to revenue acceleration | ON TRACK | Product growth accelerated sequentially to 10.7%; YoY growth moderated to 33.5%. |
| Bull 3: STT-MRAM / PERSYST growth | ON TRACK | Teledyne broadens the aerospace channel; stand-alone STT sales are undisclosed. |
| Bull 4: DoD contract validates supply-chain positioning | ON TRACK | Legacy sustainment winds down as expected; the defense relationship remains relevant. |
| Bull 5: NOR Flash conversion as a multi-year tailwind | AT RISK | Scarcity has brought initial substitution demand; repeat economics remain untested. |
| Bull 6: UNISYST launch and expanded addressable market | AT RISK | H1 2027 prototype window still leaves qualification and production volume open. |
| Bull 7: $100M revenue ambition in 3–5 years | ON TRACK | Higher near-term sales support the path; the ambition is not an annual guide. |
| Bull 8: $40M defense subcontract builds a compounding moat | ON TRACK | Continued recognition supports the signed-work case; no follow-on award announced. |
| Bull 9: Microchip manufacturing supports future growth | AT RISK | Qualification milestones are clearer but do not restore the earlier FY27 revenue premise. |
| Bull 10: Transportation / rail expands the use cases | ON TRACK | Neutral this quarter; prior design-ins remain open without a new shipment milestone. |
| Bear 1: Microcap liquidity and ownership limitations | EMERGING | The opening gap and partial reversal highlight execution risk in the stock; no ownership conclusion. |
| Bear 2: Lumpy licensing and services revenue | MATERIALIZING | Revenue held near the prior quarter; mix and milestone variability remain. |
| Bear 3: Product margins constrained to the 40s | MATERIALIZING | 43.2% missed our recovery assumption; pricing is the next test. |
| Bear 4: Patent-litigation expense overhang | MATERIALIZING | Expense breached the prior expectation; 2027 relief is a forecast, with the case still open. |
| Bear 5: Cash conversion and investment burden | MATERIALIZING | Improved operating cash still did not fund investment. |
| Bear 6: MRAM cost premium limits NOR conversion | EMERGING | Supply scarcity can win sockets even without closing the cost gap. |
| Bear 7: Manufacturing continuity after Chandler ownership change | EMERGING | No post-2028 access agreement was announced; second-site qualification remains important. |
| Bear 8: Export-control compliance and settlement | NEW → MATERIALIZING | The $2.715M provision makes the potential cost visible; settlement remains open. |
Overall: Q3 met our adoption and service tests and delivered the promised CXL demonstration. It missed our product margin and legal expense tests. The thesis still rests on shipping existing products and executing signed work, but it now takes better pricing and lower costs to turn that growth into a comparable return.
What changes the view: Q4 product margin near our 45% estimate with revenue inside guidance would support the partial recovery we model. A failure to improve from Q3’s 43.2%, another legal cost overrun or a larger regulatory payment would reduce the valuation. A credible decline in legal costs accompanied by operating cash funding capital investment would justify higher conviction.
Action: Maintain Outperform at $14.43 with 5/10 conviction and $17–20 of 12-month fair value. The lower entry price preserves expected excess return, but the risk of a decline toward $9.62 argues against increasing exposure before the pricing and cash results are visible.