Amazon is moving ahead with a plan to invest up to $50 billion to expand AI and high-performance computing infrastructure across AWS Top Secret, AWS Secret, and AWS GovCloud (US) Regions. The buildout, beginning in 2026, will add nearly 1.3 GW of dedicated AI and HPC capacity for federal agencies. The company aims to give government customers immediate access to advanced compute, accelerated networking, and a wider suite of AI services to support national security, scientific research, and industrial innovation missions.
The investment includes new data centers equipped with AWS Trainium accelerators, NVIDIA AI infrastructure, and access to Amazon SageMaker, Amazon Bedrock, Amazon Nova, Anthropic Claude, and leading open-weight foundation models. By merging AI with large-scale simulation and modeling, agencies will be able to run workflows that previously took weeks or months in a matter of hours. Defense, intelligence, healthcare, and energy missions stand to benefit from real-time pattern detection, rapid scenario generation, and AI-driven analysis of satellite imagery, sensor data, and decades of historical records.
AWS framed the plan as part of a broader federal push toward secure U.S.-based AI infrastructure under the Administration’s AI Action Plan. The company also emphasized its long history supporting classified missions, having launched AWS GovCloud in 2011, the AWS Top Secret Region in 2014, and the AWS Secret Region in 2017. More than 11,000 government agencies currently run workloads on AWS’s secure cloud.
• Up to $50 billion investment beginning in 2026
• Adds nearly 1.3 GW of AI and HPC capacity across Top Secret, Secret, and GovCloud (US)
• Expands access to Amazon SageMaker, Bedrock, Nova, Claude, open-weights models, and NVIDIA infrastructure
• Includes AWS Trainium chips and advanced HPC networking
• Enables rapid AI-accelerated modeling, simulation, and mission analytics
• Supports national security, autonomous systems, cybersecurity, drug discovery, climate and energy research, and healthcare
• Builds on more than a decade of U.S. government cloud innovation: GovCloud (2011), Top Secret-East (2014), Secret Region (2017), and multiple regional expansions through 2025
“Our investment in purpose-built government AI and cloud infrastructure will fundamentally transform how federal agencies leverage supercomputing,” said AWS CEO Matt Garman. “We’re giving agencies expanded access to advanced AI capabilities that will enable them to accelerate critical missions from cybersecurity to drug discovery.”
🌐 Analysis:
AWS is doubling down on secure AI infrastructure just months after U.S. agencies accelerated procurement of AI-ready compute and after major public-sector wins involving Bedrock and Trainium-based systems. This move positions AWS against competitors such as Microsoft Azure Government and Google’s expanding federal cloud portfolio, especially as U.S. agencies build sovereign AI stacks for defense, intelligence, and scientific applications. The scale—1.3 GW—is notable, signaling that hyperscale-class AI factories for government workloads are now moving into multi-gigawatt territory.
