OpenAI expands, flags security issue
OpenAI has secured a permanent London office in King’s Cross with capacity for roughly 544 workers and says it will open next year, signaling a larger UK presence for research and operations. The company also disclosed a security incident tied to a compromised third‑party developer tool (Axios) and said user data was not accessed while it rotates certificates and tightens app certification processes. (itpro.com) (enterpriseai.economictimes.indiatimes.com)
OpenAI is expanding in London while dealing with a separate security incident tied to a compromised outside software tool. (cnbc.com) (openai.com) The company said on April 13 that it has signed a lease for its first permanent London office in King’s Cross, with room for 544 workers and an opening planned for 2027. The site covers about 88,500 square feet in Regent Quarter, across Jahn Court and the Brassworks Building. (propertyweek.com) (cnbc.com) OpenAI currently has about 200 staff in London across research, engineering, customer support, policy, communications, marketing and sales. Phoebe Thacker, the company’s London site lead, said the new office gives OpenAI more space to keep building in the city. (independent.co.uk) (cnbc.com) The London move follows a setback elsewhere in the United Kingdom. CNBC reported that the office announcement came days after OpenAI paused plans for its United Kingdom “Stargate” data center project, a separate infrastructure effort backed by the British government. (cnbc.com) The security issue is a different story: it involved a software supply chain problem, where trusted code used by developers is tampered with before it reaches a company’s systems. OpenAI said a compromised version of the JavaScript library Axios was downloaded and executed in a GitHub Actions workflow used in its development process. (openai.com) (cnbc.com) OpenAI said the Axios package was compromised on March 31 as part of a broader industry attack that it believes was linked to North Korea, and that the issue affected the process used to certify that its macOS applications are legitimate OpenAI software. The company said it found no evidence that user data, passwords, systems, intellectual property or software releases were compromised. (openai.com) (tech.yahoo.com) In response, OpenAI said it rotated all certificates tied to the affected workflow, rebuilt impacted systems, audited recent releases and began tightening how its macOS app certification process works. It also urged users to update to the latest version of the ChatGPT macOS app. (openai.com) (cnbc.com) Taken together, the two announcements show OpenAI adding office capacity in one of Europe’s biggest artificial intelligence hubs while hardening the internal software pipeline behind its products. In London, the next marker is 2027, when the King’s Cross office is due to open. (cnbc.com) (propertyweek.com)
