Is MRAM a Buy or a Sell? The Bull and Bear Case (2026)
Last updated July 2026
Short answer
Both cases are real, which is why the question is contested. The bull case for Everspin Technologies (MRAM) rests on Defense and space work that sells process, not just parts: In late April 2026 Everspin executed a $40 million agreement covering roughly two and a half years, acting as subcontractor on an existing prime contract to supply Toggle MRAM process technology, engineering and foundry services to United States Defense Industrial Base customers. The bear case rests on the most consequential open item is litigation, and it is a patent fight rather than a shareholder suit: Avalanche Technology filed a patent infringement complaint against Everspin in the US District Court for the District of Delaware and a parallel Section 337 complaint on January 28, 2026, amended February 5, and the International Trade Commission voted to institute Investigation No. Analysts covering it publish targets from $19.00 to $38.00 against a $18.88 price, so even the professionals disagree by 74% of their own average. We do not give a verdict. What follows is each case at full strength, so you can decide which set of assumptions you actually believe. Walnut is not an investment adviser.
Everspin Technologies, based in Chandler, Arizona, builds magnetoresistive random access memory and the tunnel magnetoresistance sensors that share its magnetic process. Its memory stores each bit in a magnetic tunnel junction rather than a charge trap, so data survives a power cut, writes complete in nanoseconds, endurance is effectively unlimited, and radiation does not corrupt the cell. Those properties put the parts where flash struggles: write buffers in industrial automation controllers, smart meters and energy management gear, transportation systems, RAID cache in storage arrays, and aerospace and defense hardware where a boot image cannot be allowed to fail. Two product families carry the business, the older Toggle MRAM and the newer spin-transfer torque line sold as PERSYST, alongside a portfolio of more than 700 patents and applications. Manufacturing is split: Everspin buys standard CMOS wafers and adds its own magnetic layers at a 200mm fab it operates in Chandler, while GlobalFoundries produces full-flow 300mm STT-MRAM wafers under a long-standing agreement that grants GlobalFoundries exclusivity on certain nodes for a period after qualification. The financial picture in August 2026 is a company emerging from an industrial downcycle with two new revenue lines attached. June-quarter revenue of ~$18.7 million was the highest in Everspin's history, up ~42% from ~$13.2 million a year earlier and above the company's own ~$15.5 million to ~$16.5 million guidance, with product sales of ~$15.3 million and ~$3.4 million of licensing, royalty and engineering revenue reflecting the first work under a $40 million agreement with a US prime contractor. Gross margin came in near 53.9%, and non-GAAP net income was ~$2.9 million even as the GAAP loss widened to ~$3.6 million. Cash stood at ~$43.9 million against ~$2.8 million of debt, so nothing about the balance sheet is pressing. Valuation is where the debate sits: ~$62 million of trailing revenue supports a market value near $460 million, roughly 7.4 times sales, on a stock that has traded between $5.76 and $51.50 inside twelve months and now has a 2x leveraged single-stock ETF written on it.
The bull case: what would have to be true for $38.00
The most optimistic published target on MRAM is $38.00, +101.3% from the $18.88 price as of August 2026. Getting there needs the following to work close to its best case, not merely to avoid going wrong.
1. Defense and space work that sells process, not just parts
In late April 2026 Everspin executed a $40 million agreement covering roughly two and a half years, acting as subcontractor on an existing prime contract to supply Toggle MRAM process technology, engineering and foundry services to United States Defense Industrial Base customers. Non-product revenue jumped to ~$3.4 million in the June quarter from ~$2.1 million a year earlier as that work began, and it carries different economics from selling chips because it monetises process know-how rather than wafer volume. Alongside it, Astro Digital selected the 64Mb PERSYST STT-MRAM as the primary fail-safe boot memory for a Raven bus geosynchronous satellite mission, the kind of design win that tends to persist across a program's life.
2. A density roadmap aimed at edge AI and servers
Everspin's historical ceiling was density, and the qualification calendar is where that changes. The 64Mb STT-MRAM completed AEC-Q100 Grade 1 production qualification, the 128Mb part qualified in May 2026 and a 256Mb option in July 2026 with volume availability in the second half of the year. Above that sits the newly launched UNISYST family, spanning 128Mb to 2Gb on a standard xSPI interface running octal SPI up to 200MHz, plus CXL MRAM demonstrations moving into September and an August memorandum of understanding with MaxLinear to evaluate MRAM alongside its compression, encryption and server acceleration silicon for AI servers.
3. Onshore capacity through the Microchip agreement
An initial ten-year agreement signed in April 2026, extendable in two-year increments, puts a copy-exact 8-inch MRAM, TMR sensor and STT-MRAM line into Microchip's Fab 4 in Gresham, Oregon. First products from the collaboration are expected in the second half of 2027, so it contributes nothing to near-term revenue. What it addresses is credibility: a single company-operated fab is a hard constraint when bidding ITAR-relevant defense programs or multi-year automotive qualifications, and a second qualified US source removes it.
4. Recovery in the industrial base underneath everything else
Management attributed June-quarter product strength to industrial automation, energy management and aerospace and defense, with the industrial and energy recovery concentrated in Japan and Europe. September-quarter guidance of ~$19.5 million to ~$20.5 million implies another sequential step up from ~$18.7 million, with non-GAAP EPS guided to ~$0.10 to ~$0.15. Everspin's core markets are long-cycle and design-locked, which cuts both ways: the recovery is slower to arrive than in consumer silicon and slower to leave once sockets are won.
The bear case: what would have to be true for $19.00
The most pessimistic published target is $19.00, +0.6% from the current price. That is not a floor and not a forecast; it is roughly what one analyst thinks Everspin Technologies is worth if the risks below bite instead of the drivers above.
The most consequential open item is litigation, and it is a patent fight rather than a shareholder suit: Avalanche Technology filed a patent infringement complaint against Everspin in the US District Court for the District of Delaware and a parallel Section 337 complaint on January 28, 2026, amended February 5, and the International Trade Commission voted to institute Investigation No. 337-TA-1487 on February 27, 2026 before Administrative Law Judge Doris Johnson Hines, asserting 37 claims across four patents and requesting an exclusion order plus a cease and desist order that would bar importation of Everspin's STT-MRAM products. Everspin has stated the claims reach the STT-MRAM line only and not Toggle MRAM or sensor devices, and that it believes it has strong defenses and will contest them; a Section 337 investigation typically concludes 14 to 16 months from institution, so resolution is unlikely before roughly mid-2027, and defense costs are one reason the GAAP loss widened to ~$3.6 million in a quarter when revenue and gross margin both improved. Beyond the case, revenue is small and lumpy, a handful of industrial and defense customers and one prime-contract agreement account for a meaningful share of it, and wafer supply depends on GlobalFoundries under an arrangement that grants exclusivity on certain STT-MRAM nodes. Valuation carries its own exposure, since ~7.4 times trailing sales with GAAP losses already prices the data-center and defense roadmap converting rather than merely being demonstrated, and the trading character compounds it: the shares ran from $5.76 to $51.50 and back under $20 within twelve months, a leveraged single-stock ETF now trades against them, and realised volatility of that order is not proportional to changes in the underlying business.
The bear case deserves the same attention as the bull case, and usually gets less. If you are holding MRAM already, the question is not whether these risks exist but whether any of them has moved from possible to actually happening in the reported numbers.
Where analysts land on MRAM
3 analysts cover MRAM, with an average target of $25.67 (+36.0% against $18.88) and a split of 3 buy, 0 hold, 0 sell. Bear in mind sell-side ratings skew positive across the whole market, so that split is not a balanced vote. The full target table, the recent rating actions by firm, and how the consensus has shifted are on the MRAM forecast and price target page.
How is MRAM valued? (as of August 2026)
Snapshot for MRAM as of August 2026, sourced from Yahoo Finance and may be delayed. Valuation figures move with price and earnings; verify the current numbers with your broker before deciding.
- Revenue (TTM): ~$62M
- Q2 2026 revenue: ~$18.7M, a company record, up ~42% year over year
- Gross margin (Q2 2026): ~53.9%, versus ~51.3% a year earlier
- Q2 2026 earnings per share: ~-$0.15 GAAP, ~$0.11 non-GAAP
- Cash and equivalents: ~$43.9M against ~$2.8M of total debt
- Price to sales: ~7.4x trailing, on a market value near ~$460M
Everspin is one of those companies where the GAAP and non-GAAP lines tell different stories, and both are worth reading: stock compensation and litigation expense produced a ~$3.6 million GAAP loss in the same quarter that generated ~$2.9 million of non-GAAP income on record revenue. Trailing revenue near ~$62 million against a ~$460 million market value works out to roughly 7.4 times sales, or about 6.7 times on an enterprise value basis once the ~$41 million net cash position is netted out. September-quarter guidance of ~$19.5 million to ~$20.5 million would extend the sequential run, though a full year of the $40 million prime-contract work has not yet passed through the income statement.
How do you decide if MRAM is a buy?
Rather than asking whether MRAM is a buy in the abstract, it tends to help to answer four questions:
- Thesis: do you believe the bull case above, and is it still true today?
- Time horizon: a single stock can be volatile, so a longer horizon absorbs more of the swings.
- Position sizing: a thesis can be right and the sizing still wrong; decide how much of your portfolio one name should be.
- Overlap: check whether you already hold MRAM indirectly through an index or sector ETF before adding more.
What would change your mind on MRAM
Write the tripwires down before you need them. Deciding what would falsify your view is far harder once a position is moving against you.
- Bull case breaks if: Defense and space work that sells process, not just parts stalls in the reported numbers rather than in the narrative around them.
- Bear case breaks if: the most consequential open item is litigation, and it is a patent fight rather than a shareholder suit: Avalanche Technology filed a patent infringement complaint against Everspin in the US District Court for the District of Delaware and a parallel Section 337 complaint on January 28, 2026, amended February 5, and the International Trade Commission voted to institute Investigation No fails to materialise over several reporting periods while the drivers keep compounding.
- Neither matters if: the position has grown large enough that being wrong would damage the whole portfolio. Sizing overrides the argument.
For the full picture, see the MRAM stock guide (what the company does, the ETFs that hold it, similar stocks, and the themes it fits). In Walnut you can ask its AI about MRAM against your real portfolio and see your actual exposure before deciding.
Investing in Everspin Technologies with AI
Connect the broker you already use and ask Walnut's AI how MRAM fits what you actually hold: whether you own it already through a fund, what it would do to your concentration, and how it has tracked the S&P 500. Read-only by default, and you approve anything before it reaches your broker.
FAQ
Is MRAM a good stock to buy right now?
+
That depends on which case you find more convincing, and both are on this page. The bull case rests on Defense and space work that sells process, not just parts, with revenue (ttm) at ~$62M. The bear case rests on the most consequential open item is litigation, and it is a patent fight rather than a shareholder suit: Avalanche Technology filed a patent infringement complaint against Everspin in the US District Court for the District of Delaware and a parallel Section 337 complaint on January 28, 2026, amended February 5, and the International Trade Commission voted to institute Investigation No. Analysts covering it are spread from $19.00 to $38.00, which is itself a signal that this is genuinely contested. If you believe the thesis, the real questions become sizing and overlap rather than timing. Walnut is not an investment adviser and this is not a recommendation.
Should I sell MRAM?
+
Nobody can answer that for you, and the honest version of the question is narrower: has anything changed in the reason you bought it? The bear case on this page is the place to check. The most consequential open item is litigation, and it is a patent fight rather than a shareholder suit: Avalanche Technology filed a patent infringement complaint against Everspin in the US District Court for the District of Delaware and a parallel Section 337 complaint on January 28, 2026, amended February 5, and the International Trade Commission voted to institute Investigation No. If that risk is what you were worried about and it is now playing out, that is a real signal. If the price simply fell while the thesis held, that is a different situation entirely. The most pessimistic published target is $19.00, +0.6% from the $18.88 price, which is one analyst's downside case rather than a floor. Walnut is not an investment adviser.
What is the bull case for MRAM?
+
Defense and space work that sells process, not just parts. In late April 2026 Everspin executed a $40 million agreement covering roughly two and a half years, acting as subcontractor on an existing prime contract to supply Toggle MRAM process technology, engineering and foundry services to United States Defense Industrial Base customers. The most optimistic analyst target on MRAM is $38.00, +101.3% from the $18.88 price. That figure is only reachable if this thesis works close to its best case.
What is the bear case for MRAM?
+
The most consequential open item is litigation, and it is a patent fight rather than a shareholder suit: Avalanche Technology filed a patent infringement complaint against Everspin in the US District Court for the District of Delaware and a parallel Section 337 complaint on January 28, 2026, amended February 5, and the International Trade Commission voted to institute Investigation No. 337-TA-1487 on February 27, 2026 before Administrative Law Judge Doris Johnson Hines, asserting 37 claims across four patents and requesting an exclusion order plus a cease and desist order that would bar importation of Everspin's STT-MRAM products. Everspin has stated the claims reach the STT-MRAM line only and not Toggle MRAM or sensor devices, and that it believes it has strong defenses and will contest them; a Section 337 investigation typically concludes 14 to 16 months from institution, so resolution is unlikely before roughly mid-2027, and defense costs are one reason the GAAP loss widened to ~$3.6 million in a quarter when revenue and gross margin both improved. Beyond the case, revenue is small and lumpy, a handful of industrial and defense customers and one prime-contract agreement account for a meaningful share of it, and wafer supply depends on GlobalFoundries under an arrangement that grants exclusivity on certain STT-MRAM nodes. Valuation carries its own exposure, since ~7.4 times trailing sales with GAAP losses already prices the data-center and defense roadmap converting rather than merely being demonstrated, and the trading character compounds it: the shares ran from $5.76 to $51.50 and back under $20 within twelve months, a leveraged single-stock ETF now trades against them, and realised volatility of that order is not proportional to changes in the underlying business. The most pessimistic published target is $19.00, +0.6% from the current price, which is roughly what the stock is worth if these risks bite rather than the drivers.
What does Everspin Technologies do?
+
Everspin Technologies builds magnetoresistive RAM (MRAM) and the tunnel magnetoresistance sensors that share its magnetic process, from Chandler, Arizona.
What would have to change for MRAM to stop being worth holding?
+
Decide that in advance, because it is far harder to think clearly once a position is moving against you. The concrete tripwires here: the driver behind the bull case (Defense and space work that sells process, not just parts) stalling in the reported numbers rather than in the narrative, the risk above (the most consequential open item is litigation, and it is a patent fight rather than a shareholder suit: Avalanche Technology filed a patent infringement complaint against Everspin in the US District Court for the District of Delaware and a parallel Section 337 complaint on January 28, 2026, amended February 5, and the International Trade Commission voted to institute Investigation No) turning from a possibility into a reported fact, or the position growing large enough that a bad outcome would matter to your whole portfolio regardless of who is right.
What is Everspin Technologies, and does the MRAM ticker refer to the technology or the company?
+
Both, which causes regular confusion. MRAM stands for magnetoresistive random access memory, a memory type that stores bits magnetically and keeps them without power, and Everspin Technologies happens to trade under MRAM on the Nasdaq Global Market because it is the largest commercial supplier of that memory. Searches for the ticker frequently return technology explainers rather than company filings.
Is Everspin profitable, and does it pay a dividend?
+
Not on a GAAP basis in the most recent quarter. June-quarter results showed a GAAP net loss of ~$3.6 million, or ~$0.15 per share, alongside non-GAAP net income of ~$2.9 million, or ~$0.11 per diluted share, with the gap explained mainly by stock compensation and legal expense. Trailing twelve-month GAAP net income was around negative $2.6 million and free cash flow was slightly negative. Everspin has not paid a dividend and retains its cash.
Walnut is informational, not investment advice, and gives no verdict on MRAM. Analyst targets referenced here come from a August 2026 pull of published third-party research and change constantly. Verify current figures with your broker before acting on them.