Everspin Technologies and MaxLinear collaborate to advance memory efficiency
Companies will evaluate the use of persistent MRAM with compression, encryption and hardware acceleration technologies.
Everspin Technologies and MaxLinear have signed a memorandum of understanding (MOU) to evaluate next-generation memory architectures designed to improve the efficiency and performance of AI servers.
As AI models and inference workloads grow, server architectures must support larger datasets, expanding key-value (KV) caches and greater movement of data across systems. These demands are creating a need for technologies that combine persistent memory, acceleration and compression to improve performance, power efficiency and infrastructure utilisation.
Under the MOU, the companies will evaluate the integration of Everspin’s persistent, low-latency MRAM technology with MaxLinear’s storage compression, encryption and hardware acceleration platform.
The work will focus on metadata management, log data, write buffering, cache functions and other data-intensive workloads. The aim is to improve responsiveness, reliability and data persistence in next-generation server architectures.
“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.”
“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 utilisation, 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.”
MOU terms
The companies will evaluate:
- Technical validation of Everspin MRAM with MaxLinear’s acceleration, compression and encryption platform for AI and data-intensive server workloads
- Market development initiatives targeting hyperscale cloud, AI infrastructure and enterprise server customers
- Long-term manufacturing, supply and commercialisation opportunities
