Amazon Web Services is entering a new era—one where artificial intelligence isn’t just an innovation lever, but a multi-billion-dollar growth engine.
In its first-quarter earnings update, Amazon revealed that AWS is now operating at a $117 billion annual run rate, with 17% year-over-year growth. CEO Andy Jassy attributes much of this momentum to AI, calling it a “core reinvention layer” across both Amazon and its cloud customers.
AI Everywhere—And Amazon’s Building Every Layer
Amazon is rapidly expanding its AI infrastructure. At the chip level, AWS’ new Trainium2 processor is engineered to outperform GPU-based compute in both cost and efficiency. According to Amazon, it offers up to 40% better price-performance, a key differentiator as AI workloads scale exponentially.
For organizations building generative AI applications, AWS provides Amazon Bedrock, a managed service that allows users to work with a growing catalog of foundation models. Recent updates include access to Claude 3.5 Sonnet (Anthropic), Meta’s Llama 4, and models from DeepSeek and Mistral AI—underscoring Amazon’s commitment to model diversity.
Jassy also spotlighted Amazon Nova Sonic, a speech-to-speech AI foundation model designed to create fluid, human-like voice experiences. Alongside it is Nova Act, a research-phase AI agent capable of automating multi-step browser tasks—think workflows like booking a trip or performing enterprise IT operations.
GenAI Moves From Theory to Practice
Amazon isn’t just theorizing about AI—it’s embedding it across every division, from logistics to Alexa+ to Prime Video. Over 1,000 internal AI projects are in development, with notable traction in agentic AI, where intelligent systems handle multi-step tasks with minimal human supervision.
The company’s AI-powered assistant Amazon Q has also gained traction in software development. A new integration with GitLab Duo aims to help developers run complex workflows directly through a conversational interface—turning what used to be a long DevOps cycle into a seamless coding experience.
Infrastructure Challenges Loom, But AI Still Scales
AWS isn’t immune to the global chip crunch. Jassy admitted that while AI demand is soaring, hardware constraints—especially in GPUs, motherboards, and supporting components—are slowing how fast AWS can add capacity.
Still, the roadmap is aggressive. New hardware is en route, including expanded deployment of Trainium2 and next-gen Nvidia instances. “We’re helping as many customers as capacity allows,” Jassy said, noting that supply issues are improving quarter by quarter.
The Long Game: On-Premises Still Dominates
Jassy reminded stakeholders that cloud adoption is still early: 85% of global IT spending is still on-premises. The opportunity for AWS, then, isn’t just about AI—it’s about cloud migration. “To fully realize AI’s potential, companies must modernize infrastructure and move their data into the cloud,” he said.
Despite near-term constraints, Amazon is playing a long game with AI and cloud. With demand growing, infrastructure expanding, and new AI agents rolling out, AWS is poised to move from a $100-billion business to something even larger—one generative model at a time.