Two tech giants, AWS and NVIDIA, have joined forces to launch AWS AI Factories, a groundbreaking infrastructure solution for artificial intelligence. The new offering, launched in December 2025, brings high-performance AI systems directly to customer data centers. This partnership builds on a 15-year collaboration between the companies.
The AI Factories combine AWS Trainium accelerators with NVIDIA GPUs and special networking to create powerful AI systems. Customers can now build AI models months or even years faster than if they tried to build systems on their own. This matters for companies that need to keep up with rapid AI advances.
AWS CEO Matt Garman says the alliance improves performance, efficiency, and scalability for AI workloads. NVIDIA CEO Jensen Huang describes a “virtuous cycle” where AI demand drives more innovation in computing. The solution also helps companies meet data rules and regulations by keeping AI systems in their own facilities.
The technical design uses modular components including NVIDIA’s MGX rack architecture, GPUs, Vera CPUs, and specialized networking equipment. The solution can connect up to 72 custom ASICs with NVLink Switch technology for optimal performance. A key innovation is NVLink Fusion technology, which connects GPUs with high bandwidth and low latency. The integration of NVIDIA Nemotron models with Amazon Bedrock enables customers to build production-scale generative AI applications efficiently.
AWS adds its own custom chips like Trainium4 and Graviton CPUs to the mix. For enterprises, this means faster AI development without lengthy procurement delays. The combined hardware works as an integrated system with optimized data flow between components.
This boosts training speed and makes better use of expensive computing resources. Companies across many industries can now access AI infrastructure that was previously only available to tech giants. The AWS-NVIDIA solution helps organizations build, train, and deploy AI models while maintaining control of their data. This collaboration addresses ethical concerns surrounding AI by giving businesses greater oversight of their AI decision-making processes.
This marks a significant shift in how enterprises can access advanced AI capabilities without building everything from scratch.
References
- https://aws.amazon.com/about-aws/whats-new/2025/12/aws-ai-factories/
- https://blogs.nvidia.com/blog/aws-partnership-expansion-reinvent/
- https://developer.nvidia.com/blog/aws-integrates-ai-infrastructure-with-nvidia-nvlink-fusion-for-trainium4-deployment/
- https://aws.amazon.com/about-aws/global-infrastructure/ai-factories/
- https://www.datacenterdynamics.com/en/news/aws-launches-ai-factory-offering-that-gives-customers-dedicated-ai-infrastructure-on-premises/
