Meta acquires AI coding startup Conductor and winds down its cloud service
Meta has acquired Conductor, the startup whose app runs teams of AI coding agents in parallel, with the team joining Meta Superintelligence Labs and Conductor Cloud shutting down on 20 November.
Meta has acquired Conductor, a young artificial intelligence coding startup whose software lets developers run several AI coding agents at the same time, in a deal that underlines how fiercely the largest technology companies are now competing for the tools that write software.
The acquisition was announced on 4 November. Axios first reported the deal, and Conductor confirmed it the same day in a post on its own blog by founder and chief executive Charlie Holtz, who wrote: “I have big news: Conductor is joining Meta.” Meta confirmed the acquisition to TechCrunch, but neither side disclosed a price or any other financial terms.
Conductor was founded in 2024 in San Francisco and came through Y Combinator’s Summer 2024 batch. Its main product only launched in July 2025, yet it grew quickly enough to raise a $22 million Series A round in March 2026, led by Matrix and Spark Capital, with backing from Y Combinator and founders from companies including Notion and Linear.
What Conductor built was not another AI model. Instead, it built a control room for other companies’ models. A developer could open parallel workspaces and set several coding agents to work side by side on different parts of a project, drawing on agents such as Claude Code, Codex, Cursor and OpenCode, which come from major laboratories including Anthropic and OpenAI. Users brought their own subscriptions and programming keys, while Conductor handled the orchestration, the isolated workspaces and the cloud sandboxes in which the agents ran.
That approach captured a shift in how software is being made. Writing a single block of code with one assistant is no longer the frontier. Teams are increasingly asking groups of agents to plan, write, test and review code in parallel, and the hard problem has become keeping those agents organised, preventing them from overwriting each other’s work, and letting a human developer review what they produced. Conductor’s pitch was that it solved the coordination problem rather than the model problem.
The timing also matters. The deal landed on the same day OpenAI announced it would acquire Pierre Computer Company, a small developer-tools firm whose code-storage technology already sits inside its Codex app. Together, the two deals point to the same layer of the software stack: the unglamorous machinery between a model writing code and a reviewer merging it into a real product. Meta, for its part, launched its own coding agent, Muse Code, in August, and has been building out its Superintelligence Labs under chief AI officer Alexandr Wang.
Holtz and chief technology officer Jackson de Campos are joining Meta Superintelligence Labs, where Holtz said he looked forward to working with Wang and with Nat Friedman, the labs’ head of product. Holtz told existing customers that the team had “big, crazy plans for the next interface” and hoped to have something new for them by early next year, though he did not describe what that product would be or whether it would resemble Conductor.
For Conductor’s existing users, the announcement also begins a wind-down. New sign-ups and app downloads stopped on the day of the announcement, and free users can no longer upgrade, although existing Pro users will keep Pro features free of charge. Conductor Cloud becomes read-only on 13 November, when customers will still be able to read old chats and export their data but will not be able to send new messages, and the cloud service shuts down entirely on 20 November, after which the company will begin deleting cloud user data.
The local Mac app has a longer runway. Conductor says it will keep receiving new models and critical fixes until early 2027, when it too will begin winding down. The company has stressed that it has never sold or trained on user data and will maintain that policy, deleting all Conductor user data when the app shuts down. The practical message for teams still using the cloud service is clear: export anything that matters before the November deadline, because the window for doing so is short.
What happens next is the open question. Meta has not said whether Conductor’s multi-agent workspace will be rebuilt inside its own coding tools, offered as a standalone product, or absorbed quietly into Superintelligence Labs’ wider work. Developers who liked running rival agents side by side will also be watching whether Meta keeps that neutral orchestration alive, or whether the tool gradually tilts towards Meta’s own models. For now, the deal is best read as a talent and product acquisition in a race where the prize is not just a smarter model, but the workspace in which teams of models do real work.
More in Technology & AI
Photo: A silicon wafer used in semiconductor chip manufacturing (file image)
Technology & AIEU to invest €525 million in chip startups in bid to close the gap with the US and China
The European Union will put €525 million into chip startups, including €440 million in grants and €85 million for a new European Chips Design Platform, as Brussels tries to strengthen a sector where its design firms hold less than 1 per cent of global revenue.
Photo: Tencent headquarters (file image)
Technology & AITencent releases Hy4 AI model with cybersecurity training limited to defensive uses
Chinese technology company Tencent has released Hy4, a new open-source artificial intelligence model whose cybersecurity training was deliberately limited to defensive work, in an effort to reduce the risk that it could help carry out cyberattacks.
Photo: India Gate in New Delhi (file image)
Technology & AIIndia launches three homegrown AI models, EKA-1 and Dhir 2B and 5B, under the IndiaAI Mission
India’s electronics and IT ministry has released three indigenous foundation models under the IndiaAI Mission: EKA-1, a 26.3-billion-parameter model for Indian languages, banking and cybersecurity, and the compact Dhir 2B and 5B models designed to run offline on mobile phones.