AWS CEO Matt Garman says copying rivals is one thing Amazon does poorly

Amazon’s annual AWS re:Invent conference offered more than product announcements and technical deep dives this year. It also delivered a rare moment of blunt self-criticism from Matt Garman, the chief executive of Amazon Web Services, who openly acknowledged a weakness at the heart of Amazon’s culture. Speaking at the event, Garman said that one of the things Amazon is “particularly bad” at is copying competitors.

Garman made the remarks during a conversation with the hosts of the Acquired podcast, where he reflected on how AWS approaches product development and long-term strategy. When asked what Amazon does not do well, his answer was unusually direct. According to Garman, Amazon struggles when it tries to be a fast follower. In his words, when the company attempts to copy what others have already built, it tends to do a poor job of it.

Instead, Garman argued that Amazon performs best when it starts from first principles and focuses on solving a specific customer problem. The key, he said, is only building products in areas where the company believes it has genuinely differentiated expertise. Simply recreating what already exists in the market rarely works for Amazon, particularly at the scale AWS operates today.

That philosophy helps explain why AWS often appears slower to react to trends in their early stages, but later pushes aggressively once it commits. Rather than racing competitors feature by feature, AWS prefers to wait until it believes it can offer something meaningfully different, whether through performance, integration, or scale. It is a strategy that has frustrated some observers in fast-moving markets like generative AI, but one that Amazon leadership continues to defend.