Oracle, one of the world's largest enterprise software companies, is bringing OpenAI's AI deep into its products. At the Oracle AI World conference in Las Vegas, the company announced it will integrate ChatGPT and Codex, OpenAI's coding tool, into its applications and databases. Oracle co-CEO Clay Magouyrk shared the stage with OpenAI CEO Sam Altman.
Four new AI products
According to the announcement, the collaboration includes four products: Oracle AI Workspace, an enterprise workspace with AI capabilities; Oracle AI Security Mesh, a managed security service built on OpenAI's cyber models; Oracle App Composer, which lets users build apps with Codex while Oracle runs them in its cloud; and an AI bundle for small and medium businesses. Oracle is aiming the products mainly at fields such as healthcare and finance, where data is especially sensitive.
Google chips in Oracle's cloud
In an equally surprising move, Oracle announced it will offer Google's TPU chips, the processors Google built to train and run AI models, in Oracle Cloud Infrastructure. According to the announcement, Oracle will be the first major cloud provider outside Google to offer them. Availability is planned for 2027 and will ramp up over several years. According to CNBC, Oracle will also offer OpenAI's Jalapeño chip and Cerebras systems, and stresses it is not moving away from Nvidia processors.
Why it matters
Oracle's moves reflect two big trends in tech. First, AI is moving from chatbots into organizations' core systems, such as finance, HR and supply chain. Second, the AI chip market is diversifying, and cloud providers want to offer customers alternatives to Nvidia. For enterprise customers using Oracle systems, it means OpenAI's AI tools will reach them inside software they already know.
Keep in mind that many of the announced capabilities will only come into use later, and availability dates for some of the chips have not been fully published.
