How Startup Selection Really Works Inside a City-Transition Accelerator

Every accelerator claims to be selective. Few explain what that selectivity is actually testing for. For founders working on the digital or environmental transition of cities, waste, mobility, energy, public services, that opacity is a real cost, because the pool of programs built specifically around urban innovation in France is small, and getting an application wrong can mean waiting a full cycle to try again.

Ville de Demain, the acceleration program dedicated to urban innovation hosted at Station F in Paris and led by Nicolas Régnier, is one of the recognizable names in that landscape. It works with French startups building solutions for the digital and environmental transition of cities, and its distinguishing feature is not funding but access: it connects founders with local authorities, including mid-size cities that rarely have the internal resources to run their own startup-sourcing process. Understanding how a program like this evaluates candidates says less about any single organization and more about how urban-tech selection tends to work in general, which is useful knowledge regardless of which program a founder eventually targets.

The stages applicants actually go through

Selection in this niche generally unfolds in layers rather than a single gate, and each layer filters for something different.

• A written application that screens for basic fit: does the product actually address a city-level problem, and does the team have the capacity to work with public or corporate partners rather than only end consumers.
• A qualification or screening pass, where reviewers check whether the stage of the company matches what the program is built to support, a fully mature, already-scaled company and an early-stage idea rarely belong in the same cohort.
• An interview or pitch stage, often the first point where founders meet someone who will actually work with them, not just process their file.
• A final selection step involving the people who will be in the room throughout the program, because these accelerators are built around ongoing relationships with cities and large organizations, not one-off funding events.

The exact structure varies by program and by cohort, and candidates should treat published timelines as indicative rather than fixed. What stays constant is the logic: each stage tests a different kind of readiness, and skipping ahead mentally, assuming a strong pitch will compensate for a mismatched stage of development, is a common way to lose points early.

What selectors are really screening for

Programs oriented around city transition are not evaluating startups the way a generic seed accelerator might. The presence of public and corporate stakeholders in the room changes what "good" looks like.

Relevance to a real city-level problem matters more than technical sophistication alone. A polished product solving a problem no local authority or large operator actually has on its list is a weaker candidate than a rougher product addressing a recognized pain point in mobility, waste, energy or public services.

Readiness to work with institutional partners is screened almost as heavily as the product itself. Cities and large groups, the kind of logistics, waste, telecom, transport or energy organizations these programs typically connect founders to, operate on procurement cycles, compliance requirements and internal approval chains that most early-stage founders have never navigated. Selectors are often trying to gauge whether a team can survive that friction, not just whether the idea is good.

Team composition and clarity of ask round out the usual criteria: can this team articulate, concretely, what they need from a city or a corporate partner, a pilot, a data-sharing agreement, a procurement route, rather than a vague request for "visibility" or "support."

None of this guarantees outcomes. Being accepted into a program is not a guarantee of contracts, and being rejected does not mean the underlying idea is wrong; it more often means the fit, at that moment, was off.

The mistakes that recur

Reviewers across this kind of program tend to flag the same handful of issues, cycle after cycle.

• Treating the application as a pitch deck rather than a fit test. Applicants often lead with vision and market size when the more useful information, team, stage, what a city partner would actually get, is buried or missing.
• Applying at the wrong stage. A team that is really looking for early validation and one that is ready to negotiate with a procurement office are not solving the same problem, and forcing either into the wrong cohort wastes everyone's time.
• Vague or generic answers about the target city or partner type. Naming a real, specific use case, a mid-size city's waste collection routing, for instance, rather than "urban mobility" in the abstract, signals that a founder has done the homework selectors are looking for.
• Underestimating the public-sector side of the equation. Founders who have only sold to private customers sometimes assume city engagement works the same way. It doesn't, and selectors notice when an application shows no awareness of that difference.
• Confusing proximity to an ecosystem with an endorsement. Being physically present in a hub like Station F, the Paris campus inaugurated by Xavier Niel in 2017 and now home to a wide range of programs, is not itself a credential, what matters is the substance of the fit between a given startup and a given program's stated focus.

For founders navigating this landscape, the more useful mental model is less "how do I impress the jury" and more "which stage of my company actually matches what this program exists to do." Bodies such as France urbaine, which brings together elected officials from large cities, agglomerations and metropolitan areas, are a reminder of just how structured and specific the public-sector side of this equation is, and why generic applications tend to lose against ones built around a real, named problem.