Station F and Anthropic: The AI Boom Is Pulling Startup Founders Back Offline

Tools built on large language models let a small team, and sometimes a single person, build a product that once required an entire engineering department. In theory, such a startup has no need for an office. But Station F's Paris campus director Roxanne Varza and Anthropic's Aiden Blake said the opposite on Thursday at the Wave by Vento panel in Turin: founders still choose to sit side by side. According to Varza, more people have been coming to the campus since the AI boom began.

One caveat up front: the available excerpt of the TNW article contains no attendance statistics, and the speakers gave no exact figures. What follows covers only what is known from the published excerpt, plus context that helps make sense of it.

What was actually said in Turin

The panel's thesis is simple and a little unexpected. AI removes geographic constraints: code is written with assistants, prototypes come together over a weekend, and distribution runs through the internet. Yet demand for physical space is not falling. In Varza's view, it is growing. Her observation concerns specifically the period after the AI boom began: the campus is getting more crowded.

This assessment comes from the operator of the venue, who sees the flow of applications and residents from the inside. That makes it an important source, but also an interested one: the campus benefits if offline stays valuable. Aiden Blake represented Anthropic, the side whose models startups use as the foundation for their products. For a lab, the community of developers and founders is both a feedback channel and a market.

Why offline beats remote work in AI specifically

The explanation can be built on several observable factors, although the speakers did not name them in the available text.

• The speed of tooling change. Models and APIs update so often that knowledge of what worked this week travels faster in a conversation at the next desk than in a chat.
• Identical competitors. When a product can be built in weeks, the advantage comes not from code but from access to customers, investors and strong early hires. All of that concentrates in hubs.
• Small teams. When a company has two or three people, every mistake is costly, while testing an idea with fellow residents on campus costs almost nothing.

Details known from the article and public data:

• Event: Wave by Vento, Turin, Thursday, 8 October 2026.
• Speakers: Roxanne Varza (director of Station F) and Aiden Blake (Anthropic).
• Main thesis: AI makes it possible to build companies from anywhere, but founders choose physical proximity.
• Venue: Station F in Paris, which by the campus's own account is one of the largest startup hubs in the world, with about a thousand startups working there.
• The available excerpt does not disclose any quantitative trend in the influx.

How founders and investors should read this

First: do not take the phrase "AI will make offices unnecessary" as a law. The words of a campus operator are not proof, but they match the logic of a market where speed and connections decide everything. If you are building an AI product, work out what you actually lack: expertise, customers or money. A hub solves exactly those problems, not the question of where to put a laptop.

Second: investors should ask venues for data, not just quotes. Application trends, the share of AI companies among residents and the lifespan of startups on campus would give a real picture. Until such statistics appear in the article, the conclusion about a "return of offline" remains an observation, not a measurement.

Third: for AI labs themselves, being present at such venues is a way to keep a finger on the pulse. Anthropic's participation in the panel shows that companies selling models want direct contact with the people building products on them. If you are a developer, this is a chance to get answers about the API and its limitations directly, without waiting in a support queue.