Amazon rebuilding retail

Amazon is not talking about adding a chatbot to search results. Andy Jassy says Amazon plans to rebuild the shopping experience “from the ground up” around artificial intelligence, which means changing the underlying retail system instead of bolting new tools onto the old one. (foxbusiness.com) That is a bigger statement than it sounds. Amazon’s retail site has been tuned for decades around keyword search, product pages, reviews, ads, and checkout flows, and Jassy is saying those pieces may need to be redesigned if artificial intelligence becomes the main way people shop. (foxbusiness.com) He made the case in Amazon’s annual shareholder letter on April 9, 2026, where he argued the company should not be “conservative” in a once-in-a-generation technology shift. Amazon used that letter to defend roughly $200 billion in capital spending tied largely to artificial intelligence infrastructure. (retaildive.com, qz.com) The spending is not just for flashy consumer features. Amazon Web Services, the company’s cloud division, needs huge numbers of data centers, chips, networking gear, and power to rent artificial intelligence systems to other businesses. (cnbc.com, aboutamazon.com) Amazon also gave investors a number they had not heard before. Jassy said Amazon Web Services’ artificial intelligence business is already running at more than $15 billion in annualized revenue, which turns a vague promise into a business line with a measurable sales pace. (reuters.com, finance.yahoo.com) That cloud number helps explain the retail push. If Amazon can use the same models, chips, and internal tools in its own store, it gets one company paying for the machinery and another company showing customers what the machinery can do. (cnbc.com, foxbusiness.com) Jassy has been laying the groundwork for this for a year. In his April 10, 2025 shareholder letter, he wrote that Amazon was building more than 1,000 generative artificial intelligence applications across the company, including shopping, devices, and seller tools. (aboutamazon.com, customerexperiencedive.com) The retail side of that shows up in small pieces already. Amazon has been testing shopping assistants and recommendation tools that answer broader questions in plain language, which is different from the old system where shoppers had to guess the right search term and then sift through pages of listings. (foxbusiness.com, customerexperiencedive.com) The harder part is that Amazon makes money from the old layout too. Sponsored products, marketplace rankings, and standard search results are deeply wired into how people discover items and how merchants buy visibility, so rebuilding shopping around artificial intelligence could reshape the economics of Amazon’s storefront, not just the interface. (retaildive.com, foxbusiness.com) Jassy’s argument is that Amazon has to disrupt its own habits before someone else does. A company that spent 30 years teaching people to type “paper towels” into a search bar is now betting that the next version of shopping starts with a full question, a synthesized answer, and fewer clicks. (foxbusiness.com, aboutamazon.com)