AWS boss explains why investing billions in both Anthropic and OpenAI is an OK conflict

AWS CEO Matt Garman defended the company’s multi-billion dollar investments in competing AI firms Anthropic and OpenAI, framing the apparent conflict of interest as a standard business practice for the cloud provider. Speaking at the HumanX conference in San Francisco, Garman stated that this strategy of 'coopetition' has been part of the company's DNA since its early days. The statement is significant as major cloud platforms become critical financial and infrastructure partners for leading AI model developers, effectively choosing the frontrunners in a capital-intensive technology race.

Garman explained that AWS has long balanced relationships where it both partners with and competes against the same companies, citing how rivals like Oracle sell their databases on the AWS cloud. This established 'muscle,' he argued, allows Amazon to manage its deep financial ties—including a reported $8 billion to Anthropic and $50 billion to OpenAI—without giving its own first-party products an unfair advantage. The investments were also a strategic necessity to match offerings from its primary cloud rival, Microsoft Azure, which already provided access to both high-profile AI models.

This dual-investment strategy positions AWS as a central marketplace for AI models rather than a backer of a single champion. Garman anticipates a future where customers use model-routing services to automatically select the most efficient model for a given task, whether for reasoning, planning, or simple code completion. Such services not only provide customers with flexibility and cost-efficiency but also create an avenue for AWS and Microsoft to integrate their own homegrown models into customer workflows, ensuring their platforms remain central to the evolving AI ecosystem.

AWS is leveraging its historical 'coopetition' playbook to transform the foundation model race from a battle of single champions into a diverse marketplace it controls, ensuring its platform remains indispensable regardless of which model provider leads.