OpenAI Partners With JPMorgan to Use Codex and Its Cybersecurity Models
Key Facts
OpenAI has partnered with JPMorgan Chase to use its Codex coding assistant and cybersecurity models, company president Greg Brockman said in an interview with The Wall Street Journal during a technology conference. Brockman said the bank would use OpenAI’s models "to protect itself," in a collaboration that puts the company’s tools inside a large, highly regulated financial institution. He said the deployment would extend beyond coding or security to include using the models in enterprise transformation and knowledge work across the company. Brockman described the bank as an institution that depends on trust, underscoring the importance of the use cases he announced for the partnership. His remarks did not provide details on the deal’s financial value or when implementation would begin.
The partnership comes as Brockman said enterprise revenue has become the majority of OpenAI’s revenue, while consumer revenue continues to grow. He said the enterprise share accounted for about 30% of total revenue in an earlier period before becoming the largest source, according to his current description. In April 2026, OpenAI said enterprises accounted for more than 40% of its revenue and that it expected to reach parity with consumer revenue by the end of 2026. These figures provide context for the shift Brockman described, from a smaller enterprise share to a majority of revenue. The deal also links that business shift to the use of the company’s products in a bank’s day-to-day operations, rather than limiting them to consumer products.
The announced scope includes Codex, which OpenAI has expanded its marketing of to companies during the year. In April 2026, the company said the tool had reached 3 million weekly active users. That month, it also said it was expanding partnerships with global consulting firms to accelerate enterprise adoption of Codex, and launched Codex Labs to place its specialists directly inside customer organizations to help integrate the technology. The JPMorgan partnership fits that path, combining a coding tool with cybersecurity applications and enterprise-wide knowledge work. Brockman did not specify the number of bank users expected or which divisions would deploy the tools.
JPMorgan already has a broad base of AI tool use, according to its 2025 annual report. The bank said more than 90% of its engineers use AI coding assistants, and more than 65,000 employees in its commercial and investment banking business actively use LLM Suite. It also said in its 2024 shareholder letter that it had made LLM Suite available to more than 200,000 employees worldwide, and that coding assistants had improved software engineers’ productivity by as much as 10–20% across the software development lifecycle. The figures show that the new agreement comes at an institution with existing AI use, while the announcement adds Codex and OpenAI’s cybersecurity models. Brockman did not say how the tools would integrate with the bank’s existing systems.
The operational context also includes the bank’s scale and technology resources: JPMorganChase describes itself as having more than 320,000 employees in 66 countries and said its 2026 technology budget is about $19.8 billion. That scale gives the enterprise transformation and knowledge work Brockman referenced potentially broad reach within the bank, but the announcement did not specify rollout stages or the spending tied to the partnership. On OpenAI’s side, Brockman said on November 4 that the company generates the majority of its revenue from enterprise sales and that he believes it will be ready to go public in 2027. He described this as potential readiness, not a confirmed IPO date. Enterprise growth and the JPMorgan partnership sit within the same context in which the company’s president discussed its path toward the public markets.