Advanced Micro Devices Acquires MEXT and Approaches $1 Trillion Market Cap in Chip Rally
Advanced Micro Devices acquired MEXT, a memory optimisation company, to improve performance for AI and compute infrastructure workloads where memory bandwidth is a critical constraint. The acquisition came as AMD stock participated in a broader semiconductor rally that briefly brought its market capitalisation close to $1 trillion.
MEXT Acquisition Targets the Memory Bottleneck
Advanced Micro Devices acquired MEXT, a company specialising in memory optimisation for compute infrastructure. Memory bandwidth and latency are growing constraints in AI training and inference workloads: as model sizes increase, the ability to move data between processing units and memory quickly becomes a bottleneck that limits throughput. MEXT's technology addresses this problem, which is directly relevant to AMD's ambitions to compete with NVIDIA in the AI accelerator market.
AMD's MI300X and MI350 GPU accelerators use High Bandwidth Memory (HBM) and have attracted data centre customers looking for alternatives to NVIDIA's H100 and B200 platforms. Improving memory performance at the software and system level complements AMD's hardware investments and could differentiate its accelerators for specific workloads.
Chip Rally and $1 Trillion Threshold
AMD participated in a significant chip sector rally that added roughly $2 trillion in combined market capitalisation across Micron, Intel, and AMD. At its peak, AMD's market cap approached but did not breach $1 trillion, a milestone that would place it alongside NVIDIA in the top tier of semiconductor companies by market value. The stock subsequently pulled back from its record high as investors took profits.
AMD's rally reflects genuine business momentum: the company has gained market share in data centre CPUs through its EPYC processor line and is expanding in AI accelerators. Its roadmap for Instinct GPU accelerators and continued EPYC server processor leadership give it a credible path to sustained revenue growth in the AI infrastructure cycle.
Consumer CPU Controversy
Separately, AMD faced consumer backlash after removing memory overclocking profile support (XMP/EXPO) from some of its consumer CPU lines, a decision that frustrated enthusiast users who rely on this functionality for performance optimisation. The change is unlikely to affect AMD's financial results materially, as the consumer CPU business is smaller and less profitable than its data centre and AI segments, but it generated negative press within the enthusiast community.
Sources
Frequently asked questions
How does AMD compete with NVIDIA in AI accelerators?
AMD competes with its Instinct series GPU accelerators (MI300X, MI350) which use High Bandwidth Memory and can handle large AI model inference at scale. While NVIDIA holds a dominant position, AMD has attracted enterprise customers looking for alternatives, particularly for inference workloads where AMD's price-performance profile is competitive.
Why does memory optimisation matter for AI compute?
As AI models grow larger, the bottleneck shifts from pure compute power to how quickly data can be moved between memory and processing units. Memory bandwidth limits how fast a GPU can feed data to its computational cores. Software and hardware optimisations that improve this throughput, such as what MEXT developed, can meaningfully improve AI workload performance.
Informational only, not investment advice. Sentiment reflects news exposure, not a buy/sell recommendation or price forecast. Do your own research and consult a licensed professional.
One story is a data point. The pattern is the edge.
Reading one story at a time, you miss how the news adds up. Track AMD free and TradeTidings rolls every future headline into one clear positive, neutral or negative read, and alerts you the moment it turns.