Everspin Technologies, Inc., a developer and manufacturer of MRAM solutions, and MaxLinear, Inc., a provider of connectivity and storage acceleration technologies, have announced a memorandum of understanding to evaluate next-generation memory architectures designed to improve efficiency and performance in AI servers. The companies will evaluate the use of Everspin’s persistent, low-latency Magnetoresistive RAM, or MRAM, technology with MaxLinear’s storage compression, encryption and hardware acceleration platform.

The work will focus on metadata management, log data, write buffering and cache functions, along with other data-intensive workloads where persistent memory could improve responsiveness, reliability and data persistence.

AI models and inference workloads are creating increasing demands on server architectures as data sets grow, key-value caches expand and greater volumes of data move throughout computing systems. The companies are evaluating architectures that combine persistent memory, acceleration and compression to improve system performance, power efficiency and infrastructure utilization.

Everspin Technologies MRAM memory device | Electronics Industry Monthly

Everspin Technologies MRAM memory device. Image courtesy of Everspin Technologies.

“AI is forcing system architects to rethink where persistent memory fits in the server memory hierarchy,” said Sanjeev Aggarwal, president and CEO of Everspin Technologies. “At Everspin, we continue to advance MRAM for applications where speed, persistence and endurance need to work together. We look forward to collaborating with MaxLinear to bring that innovation into AI server architectures that need persistent memory closer to critical data, creating a path for next-generation MRAM-based solutions.”

MRAM is a non-volatile memory technology that retains stored data when power is removed. Everspin is targeting applications where persistence, low latency and endurance are important to system performance. Combining persistent MRAM with MaxLinear’s acceleration, compression and encryption capabilities could allow AI server architectures to reduce data movement while positioning persistent memory closer to frequently accessed or critical data.

“The rapid growth of AI is driving the need for smarter ways to manage, move and access data across servers,” said Vikas Choudhary, senior vice president, connectivity and storage at MaxLinear. “Combining acceleration, compression, and persistent memory creates an opportunity to improve resource utilization, reduce data movement and deliver more predictable performance for demanding AI workloads. Our collaboration with Everspin is focused on evaluating innovative architectures that can help customers scale efficiently as AI deployments continue to expand.”

Under the MOU, Everspin and MaxLinear will conduct technical validation of MRAM with MaxLinear’s acceleration, compression and encryption platform for AI and other data-intensive server workloads. The companies will also evaluate market development opportunities targeting hyperscale cloud, AI infrastructure and enterprise server customers, along with potential long-term manufacturing, supply and commercialization opportunities.

About Everspin Technologies, Inc.

Everspin Technologies, Inc. is a Chandler, Arizona-based semiconductor company specializing in Magnetoresistive RAM, or MRAM, and Spin-transfer Torque MRAM technologies. The company designs, manufactures and commercially supplies discrete and embedded non-volatile memory solutions for applications where data persistence, integrity, low latency and endurance are critical. Everspin technologies are used across data centers and AI infrastructure, industrial automation, aerospace and defense, automotive, smart energy, utilities and other mission-critical applications. With approximately 20 years of MRAM production experience, the company reports more than 200 million MRAM units shipped, approximately 2,000 customers and more than 700 patents and pending patent applications. Everspin generated $55.2 million in total revenue in fiscal 2025 and is publicly traded on Nasdaq under the symbol MRAM. For more information, please click here.

About MaxLinear, Inc.

MaxLinear, Inc. is a Carlsbad, California-based semiconductor company developing digital, high-performance analog and mixed-signal integrated circuits and software solutions for communications, connectivity and data infrastructure applications. Founded in 2003, the company’s technologies support broadband access, wired and wireless connectivity, 4G and 5G infrastructure, data center and optical interconnects, enterprise networking and storage, power management and industrial applications. Its portfolio includes high-speed optical and electrical data center connectivity, storage acceleration, RF and wireless technologies, Ethernet and Wi-Fi connectivity, broadband system-on-chip solutions, power management and interface products. MaxLinear operates across 16 countries and employed approximately 1,115 people at the end of 2025, including 786 employees in research and development. The company generated $467.6 million in fiscal 2025 net revenue and is publicly traded on Nasdaq under the symbol MXL. For more information, please click here.

Everspin Technologies, Inc.

(Editor’s Note: All trademarks mentioned in this article, including company names, product names, and logos, are the property of their respective owners. Use of these trademarks is for informational purposes only and does not imply any endorsement.)

Molly Bakewell Chamberlin
Latest posts by Molly Bakewell Chamberlin (see all)
Tagged