At a critical moment when AI competitions are entering deeper waters, Amazon is demonstrating its strategic determination through organizational restructuring. On December 18, CEO Andy Jassy announced to all employees that Peter DeSantis, a senior executive from AWS, would be appointed as the head of a new AI organization, overseeing Amazon's layout in three core areas: large models, self-developed chips, and quantum computing.

DeSantis has been with Amazon for 27 years, serving as a Senior Vice President at AWS for 8 years. He is one of the key architects behind the cloud infrastructure that supports about one-third of global internet traffic. Appointing him to lead the newly established AI department highlights Amazon's strategic logic of viewing AI as the "next evolution of the cloud."

The new organization will directly be responsible for the development and collaborative optimization of Amazon's self-developed large model series Nova (including the recently launched Nova 2 at the re:Invent conference), AI-specific chips (such as Trainium and Inferentia), and quantum computing technologies. In an internal letter, Jassy clearly stated: "We want to free Peter's energy, innovation cycles, and leadership to focus on these new areas, achieving end-to-end optimization of models, chips, cloud software, and infrastructure."

This move comes during a period of intense AI investment by Amazon:

- Last month, AWS announced a $5 billion investment in AI infrastructure for the U.S. government;

- It is reported that the company is negotiating a $10 billion strategic investment with OpenAI;

- Previously, it invested $8 billion in Anthropic, becoming its largest external shareholder.

Although Amazon started developing its own models later than OpenAI and Google, its strategy is shifting from "technology catch-up" to "ecosystem integration" — building an AI enterprise service loop centered on AWS through four drivers: cloud, chips, models, and capital. DeSantis' appointment is the organizational guarantee for this strategy.

In Jassy's view, true AI advantages do not lie in the performance of a single model but in the efficiency of full-stack collaboration. When the Nova model runs on Trainium chips, is scheduled by the AWS cloud platform, and is supported by quantum computing research for future computing power, Amazon's AI moat may shift from "scale" to "systemic efficiency."

As DeSantis, a veteran of cloud infrastructure, leads the new AI force, Amazon's AI strategy has become clear: not to create the most impressive demos, but to build the most stable enterprise-level AI infrastructure. This may be the key for Amazon to overtake in the B-side market.