AWS CEO Matt Garman says Amazon’s decision to invest in both OpenAI and Anthropic—two competing AI model companies—reflects a conflict AWS has experience managing. Speaking at the HumanX conference in San Francisco, Garman framed the issue as a familiar aspect of cloud strategy, where cloud providers often partner with companies that may also compete with them.
Amazon’s Investments in OpenAI and Anthropic
The context is Amazon’s $50 billion investment in OpenAI, following a long partnership that includes an additional $8 billion investment in Anthropic, as reported by TechCrunch. Garman was asked about the conflict of interest involved in working closely with two competitors in the AI model space.
His response: AWS is accustomed to competition with partners because it competes with them in other areas. He argued that this is not inherently problematic because the technology stack is interconnected—meaning a cloud provider and its partners can occupy overlapping roles.
AWS’s History of Managing Partner Competition
Garman traced the approach to AWS’s early years. He said Amazon knew it “couldn’t build every cloud offering itself,” so it partnered with other companies. At the same time, AWS expected it would need to compete with partners because “technology is interconnected.” AWS therefore developed a strategy for how it goes to market with partners.
According to Garman, this strategy includes the expectation that AWS might create first-party products that compete with partner offerings. He said AWS has committed to partners that it will not use unfair competitive advantages, tying the strategy to contractual boundaries rather than avoiding competitive overlap entirely.
Garman noted that today, Amazon competing with companies that sell on its cloud is familiar. He cited Oracle as an example of a major rival that sells database and other services on AWS. This arrangement was considered unusual in 2006, when technology partners typically avoided competing with those that helped them succeed.
Cross-Investment Among Competing AI Companies
Garman’s defense sits within a broader AI investment landscape. According to TechCrunch, Amazon is not alone in managing investor overlap across competing AI firms. In February, Anthropic announced a $30 billion funding round that included at least a dozen investors also backing OpenAI. This included Microsoft, OpenAI’s primary cloud partner. The pattern suggests that cross-investment among competing AI companies is already common in the industry.
Implications for Cloud and AI Infrastructure
From a technology perspective, the core issue is less about whether competition exists and more about how cloud platforms manage it while enabling model development and deployment at scale. Garman’s comments suggest AWS views competition with AI model makers as an expected outcome of building infrastructure offerings alongside partner ecosystems.
His explanation—that AWS may compete with partners through first-party products but commits not to use unfair competitive advantages—signals a model for coexistence. Overlap is expected, while incentives are managed through contractual commitments. The extent to which these boundaries hold as AI model providers deepen cloud integration and cloud providers expand AI capabilities remains to be seen.
The investor-overlap pattern from Anthropic’s funding round suggests the industry is normalizing financial ties across competitors. If this continues, it could further blur the distinction between “partner” and “competitor” in AI infrastructure, where investment, hosting, and product roadmaps can intersect.
Garman placed his remarks within AWS’s longer history. He noted that he has worked at Amazon since 2005, before AWS launched in 2006, and connected today’s approach to early decisions about partnering because AWS could not build every offering itself. This history frames the current AI conflict discussion as an extension of established cloud operating principles rather than a new exception.
For enterprises evaluating AI platforms, the practical consideration is how cloud providers manage dependencies. If major cloud players invest in multiple competing model providers while offering overlapping infrastructure and services, companies may need to assess not only model performance but also the commercial and technical relationships between providers, partners, and underlying cloud services.
Source: TechCrunch