What happened: Anthropic said it is taking London office space sized for up to 800 people, while OpenAI has secured a London site with capacity for 544 staff that is due to open in 2027.

Why it matters: Frontier labs sell automation, but they still need physical hubs for hiring, enterprise sales, and the quiet part of the business: policy access and showing risk-averse customers they have humans in the room.

Wider context: This fits the broader pattern: AI competition is increasingly fought via procurement, security posture, and political relationships, not just benchmark charts. London is a convenient nexus of DeepMind alumni, finance, and regulators.

Background: The piece notes OpenAI recently paused a UK data-center plan tied to energy costs and regulatory uncertainty, while still expanding its UK footprint via an office. It also links Anthropic expansion messaging to rapid EMEA revenue growth.

Anthropic Expands London Office as OpenAI Locks In 544-Seat UK Hub — Implicator

Singularity Soup Take: Nothing says "the robots are here" like two robot companies racing to sign leases. London is becoming the checkpoint where frontier AI turns into contracts, regulation, and "please trust us" meetings. The future is built out of GPUs and office chairs.

Key Takeaways:

• Hiring Is Strategy: Both labs are positioning London as a serious research and commercial hub, not a token outpost. Talent access and local credibility are scarce resources when every company claims it can automate half the economy.
• Policy Proximity: A local office helps in procurement-heavy sectors like finance and government, where "we have people here" changes the tone of risk conversations, even if the model runs on servers elsewhere.
• Infrastructure Vs Presence: OpenAI can pause a data-center plan because power and regulation are hard, but still expand via real estate because influence and hiring are easier. Presence is the low-capex way to stay in the game.