Replit has opened its first office outside the United States in London, and CEO Amjad Masad used the launch to criticize Silicon Valley's monoculture. The AI coding company, founded in San Francisco ten years ago, has hired around twenty people in the British capital in just a few weeks, across sales, support, and engineering. Masad said the team could reach one hundred employees by early next year, with growth tied to AI adoption.

London mayor Sadiq Khan also spoke at the launch of the European hub. Replit is not the first US company to choose King's Cross: OpenAI, Anthropic, and other startups are in the same area, near Isomorphic Labs, Meta, and Wayve. But the London opening has a more specific meaning: Replit is pushing into enterprise distribution. The company lists customers such as PwC, Ernst & Young, and Mercedes-Benz and says it has users at 85 percent of Fortune 500 companies. In March it raised $400 million at a $9 billion valuation.

Diversity as a differentiator

The most discussed part of Masad's speech compared San Francisco and London. 'I love San Francisco. Silicon Valley has given me a lot,' he said, 'but there is one major downside: it is a bit of a monoculture. Everyone speaks the same language, you can never escape tech.' In London, he added, he meets artists, people in finance, and filmmakers. For the CEO, a more diverse working environment helps generate more diverse ideas. This is not just rhetoric: in a sector selling tools for building software, variety of use cases is a concrete advantage, especially when the audience is not only developers.

Replit is among the leaders in vibecoding, the process by which non-coders and developers build software using AI. It also has an AI agent that can deploy software applications. The need to quickly hire sales, support, and engineering staff suggests the bottleneck is no longer just model capability, but the ability to help enterprises with integration.

Open source and access to AI

Masad also addressed the role of governments. The first objective, he said, is maintaining a level playing field: we do not want to repeat what happened in the last generation of platforms, with two or three companies controlling all of AI. 'We want open source AI and accessible AI.'

This position has implications beyond rhetoric. If the model market remains concentrated, enterprises have less room for local and self-hosted architectures: the choice becomes an API integration governed by a few suppliers. If the open source ecosystem remains competitive, organizations can evaluate stacks closer to direct control of data and costs. The source does not specify Replit's infrastructure, but the emphasis on accessibility touches a sensitive point for European enterprises dealing with data residency and governance requirements. For organizations evaluating on-premise deployment, trade-offs are not limited to model quality: local service, integration, and support matter. AI-RADAR offers analytical frameworks at /llm-onpremise to compare these aspects.

Replit plans to open offices in Tokyo and New York. The choice of London as its European hub, with hiring across three key functions, shows that the next AI coding battle will also be fought on operational proximity and enterprise trust, not just on the most brilliant model.