JPMorganChase to use OpenAI models for coding, security

• Key insight: OpenAI announced it’s signed a contract with JPMorganChase to use the frontier AI model provider’s technology in software development and cybersecurity.
• Expert quote: “JPMorganChase is increasingly setting the architectural and economic blueprint for how banks adopt AI.”—Krish Subramanian, founder and managing partner at Rishidot Research

JPMorganChase signed a contract with OpenAI to use the tech company’s Codex AI agents to write, review, debug and ship code, OpenAI President Greg Brockman announced via social media. The bank will also use OpenAI models for cybersecurity.

Processing Content

The announcement is significant because JPMorganChase rarely announces vendor agreements and what the largest U.S. bank does with technology, other banks tend to follow.

In the past, JPMorganChase has acknowledged the use of other AI models, for instance its LLM Suite that gives employees access to models including OpenAI’s GPT-4 and Anthropic’s Claude. Its engineering and internal analysis teams use Claude; its developers use Github Copilot among other AI coding models.

“JPMorganChase is increasingly setting the architectural and economic blueprint for how banks adopt AI,” Krish Subramanian, founder and managing partner at Rishidot Research, told American Banker. “By treating AI as core infrastructure rather than an innovation expense, they’re potentially establishing a new baseline for technology investment, operating models, and competitive expectations across banking.”

The bank’s scale and investment “could shape the technology architectures, infrastructure requirements, and vendor ecosystems that the rest of the industry adopts,” he said.

A JPMorganChase spokeswoman said the bank uses a variety of tools, “which we have consistently said we will do. And we’ll continue to use multiple.” She declined to share terms of the contract or answer questions about which OpenAI models the bank is using besides Codex.

During the bank’s third-quarter earnings call last month, Chief Financial Officer Jeremy Barnum said it’s spending “something like $700 million annually” on AI. “If you sort of price out the amount of time we’re all spending thinking about it, you can probably get a higher number, too. So at this point, it’s not really something that you can ring fence. It’s pretty well diffused inside the whole institution, and it’s obviously like a very, very high priority.”

The benefits of AI use didn’t show up in the bank’s quarterly results, he conceded.

“I think that’s probably not that surprising because … it takes a long time to really do things that are visible at an institution of our size,” Barnum said. Some experiments with using AI to rewrite “minor legacy tools” in a couple of weeks “didn’t work perfectly, but you’re starting to see some really exciting opportunities for acceleration and productivity change,” he said. “So how long it’s going to take for that to show up, we’ll see. And obviously, some of those benefits are going to be offset with token costs, which is a whole other conversation, but we’re excited.”

Barnum also said the bank’s strategy is focused on “minimizing lock-in and maximizing optionality and the ability to manage costs. So part of the strategy is model agnosticism, harness agnosticism and ability to run both proprietary models, but also open weight and open source models, some of which we run in our data center.”

Bank regulators have not offered model risk guidance specific to generative and agentic AI, though the Office of the Comptroller of the Currency recently published a framework for examiners to make risk-based assessments of AI used in banks, the way they would any technology deployment.

The most noteworthy aspect of the OpenAI deal, according to Linas Beliunas, venture advisor at EQT Ventures, is JPMorganChase’s use of OpenAI models for cybersecurity.

“Helping employees write documents is one thing, asking a vendor to help secure a bank means trusting it with a very different class of work,” he said. “This is a much bigger ambition than giving staff a chatbot.”