Peter DeSantis: Architect of Amazon’s AI and Silicon Future

2026-02-05 · Source: AI Magazine · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure, Emerging Technologies & Innovation · Depth: Novice, short

Summary

Peter DeSantis, Senior Vice President at Amazon, is a pivotal figure in the company's AI, custom silicon, and quantum computing initiatives, leading the infrastructure development for next-generation intelligence. Ranked No. 9 in AI Magazine's Top 100 Leaders 2026, DeSantis joined Amazon in 1998 and spearheaded the launch of Amazon EC2 in 2006, establishing the foundation for AWS. He also led the 2015 acquisition of Annapurna Labs, which became central to Amazon's custom silicon strategy. Currently, he oversees the Global AI Infrastructure division, focusing on breaking the cost curve of model training through aggressive deployment of Trainium and Inferentia chips and the buildout of liquid-cooled AI regions. His vision integrates AI models, custom silicon, and quantum computing as mutually reinforcing technologies.

Key takeaway

For CTOs and VPs of Engineering evaluating AI infrastructure investments, Amazon's strategy under Peter DeSantis highlights the critical advantage of vertical integration. Your teams should consider developing custom silicon and specialized data center solutions to reduce reliance on third-party GPUs, improve performance-per-dollar, and accelerate AI model development, rather than solely depending on external hardware providers.

Key insights

Amazon's AI strategy is driven by in-house custom silicon and infrastructure to reduce costs and accelerate innovation.

Principles

• Identify deepest system bottlenecks.
• Invest early in in-house capabilities.
• Integrate reinforcing technologies for compounding gains.

Method

Amazon's approach involves developing custom silicon (Trainium, Inferentia) and advanced data center architectures (liquid-cooled regions) to optimize performance-per-dollar and overcome thermal/power constraints for large-scale AI model training.

In practice

• Deploy custom silicon for cost-effective AI.
• Utilize liquid cooling for frontier AI systems.
• Integrate hardware and software development.

Topics

• AI Infrastructure
• Custom Silicon
• AI Accelerators
• Quantum Computing
• Cloud Computing

Best for: Investor, CTO, VP of Engineering/Data, AI Architect, Director of AI/ML, Tech Journalist

Related on AIssential

• Nvidia to sell 1 million chips to Amazon by end of 2027 in cloud deal — The AI infrastructure race drives massive GPU and chip stack acquisitions by cloud providers like AWS.
• Amazon Web Services Creates Forward Deployed Engineer Organization — AWS commits \$1 billion to FDEs for direct AI integration support for business customers.
• OpenAI and Amazon announce strategic partnership — OpenAI and Amazon partner to integrate advanced AI models and infrastructure, expanding enterprise AI capabilities and compute capacity.
• Google’s Cloud Next AI Highlights — Google's multi-layered AI strategy integrates hardware, compute hosting, and agent layers for a full-stack advantage.
• AWS funnels $1B into forward deployed engineering hub — Hybrid human-AI "forward deployed engineering" accelerates enterprise AI adoption and long-term innovation.
• Why Everyone Is Moving Away from NVIDIA — Hyperscalers are vertically integrating AI infrastructure, developing custom silicon and energy solutions to overcome GPU bottlenecks.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI Magazine.