AMD Acquires MEXT to Ease AI Memory Bottlenecks as Stock Climbs 265% in 12 Months
Updated
Updated · Barchart · Jun 24
AMD Acquires MEXT to Ease AI Memory Bottlenecks as Stock Climbs 265% in 12 Months
3 articles · Updated · Barchart · Jun 24
Summary
June 15 marked AMD’s acquisition of MEXT, a memory-technology company whose software pre-fetches data from flash into DRAM before GPUs or CPUs request it, reducing latency in AI workloads.
That approach aims to cut one of AI infrastructure’s biggest constraints: idle accelerators waiting on memory. By masking flash-to-DRAM delays, MEXT could lower DRAM needs, expand usable memory capacity and reduce data-center costs.
The deal adds to AMD’s AI momentum after Q1 fiscal 2026 revenue rose 38% to $10.25 billion, data-center sales jumped 57% to $5.8 billion and free cash flow reached a record $2.57 billion.
Investor enthusiasm has already driven AMD shares up 265% over the past 12 months, while management says demand for its MI450 AI GPU is running ahead of what it had expected for 2027.
Can AMD's smart memory acquisition outmaneuver the custom AI chips being built by its biggest customers like Meta and Google?
AMD's AI ambitions depend on a single supplier. Is a manufacturing bottleneck the biggest threat to the AI revolution?
Is the AI hardware war shifting from raw processing power to conquering costly data traffic jams inside servers?
AMD’s $120B AI Ambition: How the MEXT Acquisition Redefines Memory Optimization and Data Center Leadership
Overview
On June 15, 2026, AMD acquired MEXT, a company known for its AI-driven Predictive Memory Engine, to address a critical bottleneck in AI and high-performance computing. By bringing in MEXT’s skilled team and advanced memory optimization technology, AMD aims to strengthen its data center solutions and better manage memory-intensive AI workloads. Integrating MEXT’s technology expands AMD’s capabilities and complements its existing suite of processors, accelerators, networking, and software, positioning AMD to deliver more efficient and scalable AI infrastructure for the next generation of applications.