跳到正文
Silicon Republic· Tim Barnwell·· 17 天前AI 評分57

Meta 收購 AI 編碼初創公司 Conductor

Meta acquires AI coding start-up Conductor

AI 導讀

Meta 收購 AI 編碼初創公司 Conductor,交易條款未有披露。Conductor 平台讓用戶從 Mac、iPhone 或雲端運行多個 AI 編碼智能體,並行使用多個 AI 實例。Conductor 表示平台會如常運作,但 Conductor Cloud 將於未來數週關閉。

正文

The platform offers users the ability to run ‘a team of coding agents’ from a Mac, iPhone or the cloud.

Meta has acquired AI coding start-up Conductor in the tech giant’s latest move to enhance its AI capabilities and offerings.

Conductor, which was founded in 2024 by CEO Charlie Holtz and CTO Jackson de Campos, offers users the ability to run “a team of coding agents” from a Mac, iPhone or the cloud, according to its website.

Its platform uses first-party agents of Claude Code, Codex, Cursor and OpenCode to allow developers to run multiple AI instances in parallel.

Terms of Meta’s acquisition, which was initially reported by Axios yesterday (4 November), were not disclosed. In March, Conductor raised $22m in Series A funding from Spark and Matrix.

The start-up said its platform, launched in July 2025, “will keep working as normal” following the Meta acquisition, with the exception of Conductor Cloud shutting down over the coming weeks.

A blogpost on the deal also detailed a timeline and instructions for active and prospective users of the coding platform.

“I am very excited to share that Conductor is joining Meta,” Holtz wrote on X. “We have some big, crazy plans for the next interface.”

Big AI companies have been keen to buy smaller businesses and platforms that offer useful AI development tools. Yesterday, Meta rival OpenAI said it would acquire Pierre, the company behind “a family of standout open-source developer tools”.

In March, Meta acquired Moltbook, the platform for AI agents. However, it was forced earlier this year to abandon a $2bn deal for the Chinese AI start-up Manus after intervention from China’s government blocked the acquisition for reasons of national security.

Meta’s Muse Spark Open AI model was unveiled last month. Described as the “most capable” US open-weight model to date, the multimodal offering has 816bn total parameters, with 42bn active parameters and a 1m-token context window to assist in coding, agentic work and other complex, multi-step reasoning tasks.

Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.

來源:Silicon Republic · siliconrepublic.com