WorldIndustrialWeekly Analysis: Everspin MRAM, NOR Flash Price Stress, CXL Memory Tiers, Space and India

A tighter discrete non-volatile memory market is moving from supply-chain anecdote to boardroom decision. When TrendForce estimated that NOR flash contract prices rose by 100% to 120% in the first half of 2026, Everspin Technologies president and CEO Sanjeev Aggarwal described a practical response: treat MRAM as a performance and availability hedge. In remarks reported by EE Times, he said, “If you care for performance, I think that MRAM is the only solution we have,” positioning the company’s magnetic memory as an alternative at the edge, in satellites, in server infrastructure and in industrial control paths where long lead times and high reliability reshape sourcing choices.

For manufacturing, transportation and energy-system buyers, the story is not simply which memory vendor is winning. It is how allocation, price and qualification risk move component substitution from the purchasing desk to the engineering architecture. The same constraints that push customers toward cross-reference searches are also pulling Everspin into discussions around CXL memory pools, aerospace boot memory and AI-server workloads.

From NOR Flash Availability to MRAM Substitution

Aggarwal tied the near-term opportunity to customers checking MRAM for performance, reliability and supply. “Now people cannot buy NOR flash,” he said. “Most of the fabs in Asia have converted from NOR flash to DRAM. Much better profit margins. Then what is the alternative? MRAM is the only alternative for standalone or discrete products. And that is where they approach us.” He estimated that the gap left by flash shortages could be worth $3 billion in 2028 or 2029, and said Everspin has been developing a product for that market using TSMC's manufacturing process.

Cost, he argued, also became easier to justify as flash pricing rose. He said MRAM previously cost about three times as much as NOR flash, then narrowed to about 1.5 times. “If you are going to use two NOR flash chips, you might as well just use one MRAM and be more cost-effective,” he said.

The TrendForce chart dated June 2026 shows SLC NAND and NOR Flash supply, demand and pricing conditions for the second half of 2026.

Everspin specifies minus 40°C to 125°C for the parts discussed, with reliability, SRAM- and DRAM-like performance and read-write endurance cited as selection drivers.

CXL Tests a Persistent Tier Between Compute and NAND

At the Storage Networking Industry Association Developer Conference in Santa Clara, California, Everspin demonstrated MRAM modules combined with a Compute Express Link (CXL) controller. The system used CXL.mem to create a pool of persistent memory. Aggarwal said the goal was to test MRAM as an intermediate tier between DRAM, working memory, and NAND flash, storage. He contrasted DRAM with NAND access that can take hundreds of microseconds; moving data between compute and slower storage can delay AI workloads.

“Our memory is 20 to 30 nanoseconds,” he said. “You can use it for storage as well as compute, reducing the gap between NAND storage and DRAM working memory.”

He said the technology could hold data generated during AI processing before transfer to longer-term storage, including key-value caches, checkpoints and data waiting for write. Current density reaches one gigabit, but devices could be stacked to reach 8 Gb or 16 Gb. Capacity remains a limitation, though he argued it would not need to match NAND if used temporarily for intermediate data.

Everspin also has supplied IBM's FlashCore Modules and is working beyond DDR4. “Our demonstration of the CXL protocol with our existing DDR4 chip will start getting us some entry into the data center market,” he said. “And by the time we have a good foothold, we will have a 4-Gb DDR6 part ready to go right behind it.” Aggarwal estimated that development would take about three years. At the edge, he cited OpenMV as an example where Everspin replaced SRAM and added MRAM for configuration.

A GEO Boot-Memory Case and Radiation Tradeoffs

Earlier this year, Astro Digital selected Everspin's 64Mb PERSYST spin-transfer torque MRAM (STT-MRAM), part number EM064LXQADG13IS1T, as the primary fail-safe boot memory for its Raven satellite bus. Boot memory stores code needed to start a spacecraft's electronics and support system recovery.

“What Astro Digital did recently was, they took our part and sent it into geosynchronous orbit,” Aggarwal said. “So, not low Earth, but they sent those to geosynchronous orbit.”

He said MRAM's radiation tolerance helps in space, while surrounding silicon can still be vulnerable; a backup memory provides an additional safeguard if nearby circuitry is affected. Aggarwal also cited write speed and reliability as advantages over NOR flash, saying the memory can be written almost three orders of magnitude faster than NOR flash and uses less power in the applications described. Everspin lists 64-Mb, 128-Mb and 256-Mb capacities in the PERSYST family, which uses QSPI.

Aggarwal gave a general chip-count reduction example for some applications and said it should not be read as describing Astro Digital's actual satellite configuration.

AI Server Workloads Under Evaluation

The company signed a memorandum of understanding with MaxLinear to evaluate MRAM alongside MaxLinear's storage compression, encryption and hardware acceleration platform for AI servers. The work targets metadata management, log data, write buffering, cache and other data-intensive workloads, plus technical validation and market development aimed at hyperscale cloud providers, AI infrastructure developers and enterprise server customers.

“They have accelerators, so they were very interested in the technology,” Aggarwal said. “We are working with them to find the first use case.” The agreement does not announce a finished product or volume commitment.

Two-Sourcing as an Operational Hedge

Everspin produces toggle MRAM at its 8-inch facility in Chandler, Arizona, and has shipped more than 200 million units over 20 years with only a handful of returns. Its portfolio ranges from 128 kilobits to 1 Gb.

Earlier this year, the company signed a 10-year foundry services agreement with Microchip Technology to build an 8-inch MRAM and tunnelling magnetoresistance sensor line at Microchip's Fab 4 in Gresham, Oregon. The U.S. SEC filing stated first products are expected to ship in the second half of 2027.

“We signed a foundry services agreement with Microchip, which allows me to manufacture twice the amount of toggle MRAM that I could manufacture in Chandler,” Aggarwal said. The agreement is connected to a $40 million U.S. government contract awarded through Amentum Services, with Everspin as subcontractor. Aggarwal said the work will qualify the process at Microchip and provide government access to intellectual property and a second manufacturing line.

For STT-MRAM, Everspin has worked with GlobalFoundries on 12-inch wafers since 2016 and plans to tape out a product using TSMC's 16-nm process to diversify supply. The aim is two sources for each family: Chandler and Oregon for toggle MRAM; GlobalFoundries and TSMC for STT-MRAM. Aggarwal said capacity was not currently a constraint.

Competition remains. He cited South Korean company Netsol, which uses Samsung Foundry, offering a 16-Mb product, and expects a 1-Gb product from the competitor to reach the market in 2028 or 2029.

Revenue, Guidance and Risk Framing

According to its annual report filed with the U.S. Securities and Exchange Commission (SEC), Everspin reported revenue of $55.2 million for 2025, up 9.5% from $50.4 million in 2024. Aggarwal expects the company to double revenue over three to four years, with the NOR flash opportunity and demand from data centers, aerospace and defense, low-Earth-orbit satellites and industrial automation contributing to growth ambitions.

India as an Aerospace, Automotive and Rail Opening

Aggarwal said India is another potential market, particularly in aerospace and defense, electric vehicles and railways. He is seeking connections with the Defence Research and Development Organisation (DRDO), the Indian Space Research Organisation (ISRO), Indian satellite companies and Indian automotive groups. Everspin is working to expand its Indian distribution network and is exploring local partnerships, including discussions involving Indian distributors and possible outsourced semiconductor assembly and test arrangements.

He identified battery management systems, vehicle data recorders and flight control systems as possible applications. He cited existing deployments in rail systems in Europe, Japan and China, where memory stores operational data. “We are always in discussions about these potential opportunities,” he said. “There is nothing that I can talk about openly, but obviously the conversations are ongoing.”

Common Misconceptions and Buyer Checks

• Misconception: MRAM is an instant universal replacement for NOR flash. Reality from the interview: suitability depends on application, chip count, performance and reliability specifications, density and interface; Ferroelectric RAM has lower available densities, and NVSRAM can present tradeoffs in reliability and operating conditions.
• Misconception: The CXL demonstration is a shipping data-center product. Reality: It is a demonstration and entry strategy; roadmap parts remain under development.
• Misconception: The MaxLinear MoU signals volume supply. Reality: It establishes an evaluation framework, not a finished product or volume commitment.

Action recommendations: track lead times; ask distributors for cross-reference information when particular NOR flash components are unavailable; validate the temperature range minus 40°C to 125°C for mission-critical nodes; assess whether one MRAM device can replace two NOR flash chips; consider second-source manufacturing and qualification timelines; and treat space, AI-server and CXL cases as staged pilots rather than immediate broad substitution.

FAQ

What triggered the NOR flash shortage narrative?

The TrendForce report said AI-related demand absorbed much of the limited supply, and NOR flash was among the first products to face longer lead times and allocation. It also noted that suppliers had not announced major capacity expansion.

Why might MRAM be used in satellites?

Aggarwal cited radiation tolerance, write speed and reliability, with the caveat that surrounding silicon can still be vulnerable to radiation damage.

What limits the CXL memory-tier idea?

Capacity. Current density reaches one gigabit, but devices could potentially be stacked to provide 8 Gb or 16 Gb.

Which markets are near-term and long-term?

Industrial automation, aerospace and defense, data centers, low-Earth-orbit satellites and NOR flash substitution are named growth drivers; the MaxLinear AI-server work is an evaluation, while India opportunities remain potential across aerospace and defense, electric vehicles and rail.

Why did India appear?

Aggarwal said he was seeking connections with the Defence Research and Development Organisation (DRDO), the Indian Space Research Organisation (ISRO) and Indian satellite companies, while exploring automotive, rail and distribution partnerships. He cited existing deployments in rail systems in Europe, Japan and China, where memory stores operational data.

For now, the challenge is to convert MRAM's performance and reliability advantages into broader adoption across these markets. Aggarwal expects demand to grow as manufacturing efficiencies reduce cost per bit.