AMD Acquires MEXT Startup to Optimize Data Center Memory
Photo: Tom's Hardware
Quick answer
AMD acquired MEXT to implement multi-tiered memory technology, replacing expensive DRAM with NAND flash while maintaining performance through AI-driven data access optimization.
AMD has announced the acquisition of MEXT, a startup specializing in multi-tiered memory technologies for data centers. MEXT's solution enables NAND flash memory to function as DRAM, reducing hardware costs and improving infrastructure efficiency. The technology is particularly valuable amid the growth of AI models, where memory capacity is a critical performance factor.
The core of the solution is the Predictive Memory Engine—a system powered by artificial intelligence that analyzes data access patterns. Rarely used data is automatically moved from expensive DRAM to cheaper flash memory, while applications continue to interact with it as if it were RAM. This approach lowers the total cost of ownership without compromising performance.
AMD plans to integrate MEXT's technology into its data center products, including solutions for AI workloads. The acquisition will also enhance the company's expertise in memory architectures and large-scale computing. Financial terms of the deal were not disclosed, but it is expected to strengthen AMD's position in the enterprise and cloud solutions market.
Common questions
- What is MEXT technology?
- MEXT developed a memory-tiering system that uses AI to analyze data access patterns, moving rarely used data from DRAM to cheaper NAND flash. Applications perceive flash as RAM, ensuring seamless performance.
- Why does AMD need this technology?
- MEXT's technology allows AMD to offer clients more economical data center solutions, especially for AI workloads where memory capacity is often a bottleneck. This reduces total cost of ownership and enhances infrastructure efficiency.
- How will this impact the data center market?
- The adoption of this technology could reduce reliance on expensive DRAM, optimize existing hardware usage, and accelerate the deployment of large-scale workloads, including AI models. This will intensify competition among providers of enterprise and cloud infrastructure solutions.
Dzen feed: /feed/dzen.xml · RSS: /feed.xml