Oracle Embeds ChatGPT and Codex in Apps, Adds Google and Cerebras Chips to Cloud

Oracle announced on Tuesday that it has integrated OpenAI’s ChatGPT and Codex tools into its own software applications. This move allows users to build, deploy, and manage applications directly within Oracle’s database environment. The company stated that this integration makes database systems more actionable for businesses in healthcare and finance.

Oracle also expanded the variety of hardware available on its cloud infrastructure. The platform will now support Google’s tensor processing unit AI chips and OpenAI’s Jalapeño chip. It is the first time Google’s custom chips will run on a competitor’s cloud infrastructure. Oracle also confirmed that Cerebras hardware will operate on its cloud platform.

Mahesh Thiagarajan, Oracle’s president of Cloud Infrastructure, noted that his team observed OpenAI’s GPT-6 Astra model being trained at the Stargate data center campus in Abilene, Texas. He said this demonstrated the real-time benefits of using Oracle’s facilities. Greg Brockman, president and co-founder of OpenAI, stated in a press release that more businesses are choosing OpenAI because it becomes easier to use AI through existing tools and systems.

Thiagarajan addressed the ongoing scarcity of AI chips, explaining that having different accelerators benefits customers like OpenAI. When asked if this signaled a move away from Nvidia GPUs, he said no. He described Nvidia chips as the best and most fungible, while noting the AI market remains large. Sam Altman, OpenAI’s CEO, is scheduled to join Oracle co-CEO Clay Magouyrk for a keynote address at 12:40 p.m. ET on Tuesday.

The announcement follows a similar move by Amazon Web Services, which announced plans to bring Cerebras hardware into its data centers seven months ago. Cerebras stock rose about 8 percent following the news. Oracle shares increased by nearly 5 percent on Tuesday. A Cerebras spokesperson said the deal is material but the company is not disclosing the scale because it is currently in a quiet period.

Oracle also discussed its Project Jupiter data center in New Mexico, which has faced pushback from the local community. Thiagarajan said the project is one of 30 to 40 data center projects in development per quarter for the company. He said the company is building global infrastructure and is not overly concerned about the local resistance. A hearing on air quality permitting is set for Dec. 7 in Sunland Park, where the project has already broken ground.

The controversy surrounding the New Mexico project led Oracle to issue a force majeure notice to its backer, Blue Owl Capital, last month. Oracle maintains that the project is currently on schedule. Investors are awaiting Oracle’s upcoming investor day, where the company may address its fiscal year 2030 revenue targets. Bank of America analysts expect the meeting to provide clarity on major data center projects and the margin profile of cloud deals.

Oracle shares have fallen more than 30 percent this year. This decline is attributed in part to concerns over how the company plans to finance its digital infrastructure pipeline. The expansion of its hardware options and AI integrations comes as Oracle looks to secure its position in the enterprise AI market.