The company that transformed online commerce is looking to capitalize on the generative artificial intelligence boom and hopes much of it will be built on its cloud computing service.
In his annual letter to shareholders published on Thursday, Amazon CEO Andy Jassy focused on generative artificial intelligence – that is, chatbots and image generators – and where he thinks Amazon will be in the future.
“Generative AI may be the biggest technological transformation since the cloud (which itself is still in its early stages) and perhaps since the internet,” Jassy wrote. “There has never been a time in Amazon’s history when we felt there were so many opportunities to make our customers’ lives better and easier.”
Jassy said generative AI is the company’s “another set of primitives” — or “fundamental building blocks that designers can weave into any combination” — for improving the customer experience. Amazon sees three layers of generative AI and is investing in each of them, he added.
The base layer consists of developers and companies that build foundation models, Jassy wrote. Amid “sporadic” supply and cost issues with Nvidia chips, Jassy pointed to how Amazon built its own AI training chip, called Trainingand the inference system, the so-called Inference, adding that the second version of the chips released last year is “significantly more profitable than the first versions and other alternatives.” Anthropic, Hugging Face and Qualtrics are among Amazon’s chip-buying customers.
The “middle” layer of the stack contains clients building generative AI models on Amazon’s Bedrock service which allows companies like Delta Airlines and Pfizer build your own generative artificial intelligence products.
Jassy said Amazon is “building a significant number of GenAI applications” for its consumers, including its AI-powered shopping assistant, Rufus, as part of the “top-tier” tier. (The bot, for its part, has been criticized as an insignificant improvement to your shopping experience at bestThe company also builds applications on Amazon Web Services (AWS), including what Jassy called “probably the most compelling early-stage apply case for GenAI” — a recently launched coding tool called Amazon Q that writes, debugs, tests and implements software code.
Inside the company, Amazon’s robotics team is building a set of core AI models, Jassy said, including models that enable “better product identification in intricate environments” and “optimizing the movement of our growing fleet of robots.”
However, the company does not see itself as the primary developer using its technology. While Amazon is working on building its own AI applications, Jassy said the “huge majority” will be built by other companies.
“[W]The hat we build in AWS is not just an attractive application or a basic model,” Jassy said. “These AWS services, across all three layers of the stack, include a set of primitives that democratize the next, disruptive phase of artificial intelligence and will enable internal and external creators to transform virtually every customer experience we know (and invent entirely up-to-date ones that are as good as it gets). We are hopeful that much of this world-changing AI will be built on the AWS platform.”
He added that AWS and its partners offer customers running their AI services on AWS “the strongest security capabilities and greatest achievements in the world.”
In March, Amazon completed its work $4 billion investment in AI startup Anthropic — the largest-ever investment in a third-party company — boosting efforts to compete in the AI space with other industry giants like Microsoft.