AWS commits up to $50B to build AI + HPC infrastructure for U.S. government missions
If you run programs that touch national security, research, or public services, this matters. AWS is building the first AI and high-performance computing infrastructure purpose-built for U.S. government workloads across all classification levels. Construction starts in 2026, adding nearly 1.3 gigawatts of compute capacity across AWS Top Secret, AWS Secret, and AWS GovCloud (US) Regions.
Key takeaways
- First-ever AI + HPC infrastructure purpose-built for U.S. government missions.
- Nearly 1.3 gigawatts of new capacity across Top Secret, Secret, and GovCloud (US) Regions.
- Expanded access to AWS AI services and secure, scalable infrastructure to advance U.S. AI leadership.
What's being built (and where)
AWS will deploy new data centers with advanced compute and networking across AWS Top Secret, AWS Secret, and AWS GovCloud (US). The footprint is designed to handle classified and controlled unclassified workloads at scale, with isolation, compliance, and high-throughput networking baked in.
For agencies, that means more headroom for AI training, simulation-heavy analysis, and mission systems that need low latency and strong security controls.
What agencies get on day one
Federal teams will gain expanded access to a full AI stack: Amazon SageMaker for training and customization, Amazon Bedrock for model and agent deployment, Amazon Nova and Anthropic Claude, leading open-weight foundation models, AWS Trainium chips, and NVIDIA AI infrastructure. In short: more choice, more capacity, and the ability to stand up secure AI workflows without rebuilding your toolchain.
How this accelerates missions
- Modeling and simulation + AI: Move from weeks to hours using autonomous experimental steering and real-time feedback loops.
- Global security analysis: Process decades of data across hundreds of variables in real time and surface patterns as actionable insights.
- Supply chain, infrastructure, environment: Fuse fragmented datasets into a unified picture for faster planning and risk response.
- Defense and intelligence: Detect threats and draft response plans by processing satellite imagery, sensor streams, and historical patterns at scale.
- Workforce productivity: Pair analysts, researchers, and operators with AI agents to reduce manual review and speed decision cycles.
AI + HPC are converging
Federal customers are moving from traditional batch HPC to AI-accelerated discovery. Researchers and engineers can work through natural language interfaces, orchestrate expert models and agents, and validate recommendations against high-fidelity simulations. This shifts teams from "collect and wait" to "ask and iterate."
Security, compliance, and a track record that matters
This expansion builds on over a decade of government cloud milestones: GovCloud (US-West) in 2011 for security and compliance needs; the first air-gapped commercial cloud for classified workloads in 2014; accreditation across Unclassified, Secret, and Top Secret in 2017; and continued regional growth through 2025, including GovCloud (US-East), Top Secret-West, and Secret-West.
Bottom line: agencies get accredited environments, comprehensive governance controls, and proven operations at scale-so teams can focus on mission outcomes instead of managing on-prem systems.
Policy alignment
The investment supports Administration priorities for secure AI and advanced computing on U.S.-based infrastructure. For context on federal AI direction, see the Executive Order on AI from the White House here.
What to do next if you lead a federal program
- Map workloads by classification and latency needs to plan early capacity requests.
- Prioritize 2-3 high-value use cases (e.g., mission analytics, modeling + simulation, document workflows) for quick wins.
- Improve data readiness: access paths, lineage, labeling, and policies for model training and evaluation.
- Stand up a pilot with clear metrics (speed, accuracy, cost per analysis) and a governance checklist.
- Upskill your team on AI tooling and safety practices. Curated options by role are available here.
Availability and impact
New capacity begins construction in 2026 and will be available to existing and future U.S. government customers across AWS Top Secret, AWS Secret, and AWS GovCloud (US). The goal is straightforward: give federal agencies secure scale for AI and supercomputing, speed up discovery, and support America's leadership in AI-driven research, security, and public services.
Helpful resources
- Overview of AWS GovCloud (US) compliance and services: aws.amazon.com/govcloud-us
Your membership also unlocks:
