AWS to Deploy 2 Million More NVIDIA GPUs by 2028

AWS’s commitment to deploy 2 million more NVIDIA GPUs by 2028, on top of 1 million already planned, underscores the intense demand for AI infrastructure. For Asia, this signals continued pressure on regional cloud providers and data centers to scale their own GPU offerings or risk falling behind. Companies in markets like Singapore and Japan, which are actively pursuing AI development, will benefit from the increased global capacity, but local access and pricing remain key considerations. The partnership extends beyond just GPUs, with plans for NVIDIA Vera processor infrastructure and integrating NVIDIA technology into Amazon EMR and OpenSearch. This full-stack collaboration, including Amazon Robotics using NVIDIA’s physical AI platform, points to a deepening integration that could accelerate AI adoption in manufacturing and logistics across Asian economies. The scale of this investment by AWS and NVIDIA suggests a long-term bet on the foundational role of specialized hardware for AI development. The thing to watch is how this massive deployment affects the competitive landscape for AI training and inference in Asia. While AWS’s global reach benefits Asian customers, the concentration of advanced GPU capacity with a few major players could limit options for smaller regional startups or those seeking sovereign AI solutions. The US government’s plan for 100,000 GPUs also highlights the strategic importance of AI hardware, potentially influencing how other nations, including those in Asia, approach their own AI infrastructure security.