More than 900 million weekly users will be affected after OpenAI co-founder Greg Brockman took formal control of product strategy and moved to merge ChatGPT and Codex into a single product. Wired reported Brockman outlined the consolidation in an internal staff memo while he was overseeing products on an interim basis during CEO of AGI deployment Fidji Simo’s medical leave. The memo says OpenAI will embed Codex-style automation and programming functionality directly into the conversational ChatGPT interface, and the company told Wired that Simo worked with Brockman on the changes while she remains on medical leave.

OpenAI has recast product leadership around a single, agent-focused platform, according to reporting in Wired. The change formalises what had been an interim arrangement with Greg Brockman overseeing product while Fidji Simo is on medical leave. In an internal memo obtained by Wired, Brockman described a consolidation designed to "execute with maximum focus toward the agentic future, to win across both consumer and enterprise," the memo said.

What Brockman told staff

The memo, as reported, lays out a direct product consequence for both end users and developers. ChatGPT and Codex will be combined into a single experience, with Codex-style automation and programming capabilities embedded inside the conversational interface. That's a material change. An internal figure cited in the reporting said ChatGPT has grown to over 900 million weekly active users, which helps explain why OpenAI is moving to centralise those capabilities under one product umbrella.

The internal reorganisation also names specific people to run the combined platform. Thibault Sottiaux was named leader of the merged product platform that will cover consumer, enterprise, and developer surfaces. Nick Turley will shift to concentrate more on enterprise revamps while retaining responsibilities for ChatGPT. OpenAI framed the consolidation as a way to centre resources on a unified agent platform that blends conversational AI with coding and automation functionality.

The consolidation follows a pullback from peripheral projects, which the reporting links to last year’s internal "code red" directive from CEO Sam Altman. Since that directive, OpenAI stopped work on several side projects, including the video generator Sora and an OpenAI for Science initiative. The company has also spoken publicly about building an AI "super app" that centralises agentic features, and the recent moves make that strategy operational rather than aspirational.

Compute and data-centre shakeup

The product reorganisation sits alongside a separate reshaping of OpenAI’s compute and data-centre strategy. DatacenterDynamics and The Information report that OpenAI has split compute and infrastructure teams into groups responsible for data-centre design, commercial partnerships with cloud and chip vendors, and management of the facilities the company uses.

Mastheads

The split is intended to better align product ambitions with physical compute capacity.

One visible effect is a reduced expansion plan for the Abilene Stargate site. The company cut the planned expansion to 1.2GW, down from an earlier 2GW plan. Sachin Katti, who joined OpenAI in November 2025, was named to oversee the Stargate groups. In a statement he said, "AI demand is moving fast, and our compute team has built a powerful ecosystem across clouds, chips, and infrastructure to meet it."

OpenAI has broadened cloud commitments in recent months, including a deal to use 2GW of AWS Trainium chips. Reporting also flagged a reported $100 billion expansion to that cloud agreement in late February 2026. At the same time the company revised its multiyear compute spending plans to roughly $600 billion over four years, down from an earlier figure of $1.4 trillion over eight years. That rebalancing of ambition and capital suggests OpenAI is prioritising concentrated, interoperable capacity over a wider, more diffuse footprint.

Sources in the reporting say OpenAI intends to align the consolidated product roadmap with tighter infrastructure partnerships so the unified ChatGPT-Codex experience can surface automation and agent capabilities for both consumer and enterprise customers. That alignment is as much about product timing as it's about feature set, since agentic tools demand low-latency, high-throughput compute and predictable vendor relationships.

OpenAI also said it has more than half a dozen data-centre sites under development, including a partnership build with Oracle in Wisconsin where the first steel beams went up this week. The reporting doesn't specify public rollout dates, pricing, or region-by-region availability for the consolidated ChatGPT-Codex product. For customers and developers, the key unresolved questions are when the merged interface will arrive, how existing integrations will map into the new platform, and which enterprise features will follow.

The leadership reshuffle and infrastructure changes also come amid turnover at the company and intensifying industry competition. The reporting links those pressures to the broader refocus on core agentic capabilities. This practical consequence is a narrower set of priorities and a clearer centre of gravity for product decisions, with Greg Brockman now positioned as the steward of that centre.

OpenAI told Wired that Fidji Simo worked with Brockman on these changes. Simo remains on medical leave. The company has presented the moves internally as a way to concentrate resources and speed decision-making as it pursues a single platform that serves both consumer and enterprise use cases.

Related Articles

OpenAI says it has more than half a dozen data-centre sites under development, including a partnership build with Oracle in Wisconsin where the first steel beams went up this week. The company hasn't given dates for a public rollout of the merged ChatGPT-Codex product, leaving users and developers waiting for a timetable and details about integration and enterprise features.

This article was created with AI assistance.