🇫🇷 YC Spring 2025: The successful bet of 10 French startups

In the Spring 2025 promotion of Y Combinator, France is illustrated with 10 startups founded by French have been selected, a large majority based in France at the time of their selection. If the profiles are varied, a logic emerges: that of a Applied, vertical, designed to solve concrete problems in often complex sectors.

The return of the infrastructure, by AI

Blaxel With Christophe Ploujoux, Charles Drappier, Paul Sinai, opens the march. The startup builds a cloud infrastructure specially thought of for the execution of AI agents. Where AWS or Azure bet on raw power, Blaxel seeks fine optimization with persistent memory, parallel orchestration, dynamic cost management. It is part of an emerging trend that some already nicknamed the “Serverless Agent-Native”.

Other technical initiative, Nao Labs With Claire Guy, aims at the data teams. Its platform intends to modernize the development experience for data engineers, intelligent publisher, self -control, versions management, real -time collaboration. Again, a very targeted need, so far badly covered.

Accounting, audit, compliance: AI at the service of complex functions

Three projects attack functions that are still not very automated in companies.

Combinely Created by Tom Invernizy, develops a co-pilot for accountants, with automation and context assistance tools. Not a general chatbot, but a tool trained on workflows, documents and anomalies typical of the sector.

Moby AnalyticsCo founded by Thomas Rapilly, for its part, targets financial listeners. Its platform is based on specialized agents, capable of ingesting complex accounting files and extracting alerts or inconsistencies, in a format usable by an audit firm.

Finally, Probo Created by Antoine Bouchardy, intends to simplify regulatory compliance for SMEs. Its ambition is to offer a single dashboard where to follow its obligations, generate the necessary reports, automate documentation procedures, in a European context where the regulations are intensifying.

AI and health: between steering and prevention

Two other French batch startups position themselves on health, with additional approaches.

Kaelio Created by Lucas Martial, designs an operating system for hospitals. This is not a question of diagnosis, but of management of operations: care planning, resource management, team coordination. The objective is to provide health establishments with a management tool as fluid as an industrial ERP, with an additional AI layer.

At the other end of the spectrum, Lucis Founded by Max Berthelot, Max Guerois, and Baptiste Debever relies on preventive health. Inspired by the American model Function Health, it offers an offer of advanced and personalized check-ups, focused on biological data. The challenge: to give everyone a proactive vision of their state of health, beyond generic recommendations.

AI interfaces designed for teams

Nomi Co founded by Ethan Safar and Swan Beaujard illustrates this desire to build useful tools, directly exploitable in business. The startup is developing a co -pilot for sales teams. Integrated into online calls or exchanges, he suggests answers, detects objections and offers recovery areas. A time saving, but above all a skill rise.

In a more operational register, HQ flow Created by Sylvanus Mahé, is aimed at transport and logistics companies, with a series of management tools adapted to their daily lives: tours, costs, follow -up. Again, AI is present, but through discreet and integrated decision support functions.

Safety and AI: Anticipate the risks

Finally, Tinfoil Created by Tanya Verma, Sacha Servan Schreiber tackles a subject that has become strategic with the security of IA workflows. Its platform makes it possible to draw the data, to audit the models used, and to guarantee the confidentiality of treatments. A growing need for companies that handle sensitive data, in a context where AI can no longer remain a black box.

A pragmatic generation

By bringing together these 10 startups, an observation is essential: France does not seek to build “the next chatgpt”. Rather, she invests complex verticals, with a strong business issue, with tools cut for professional users. This pragmatism, coupled with a high density of technical profiles, seems to convince YC.

Especially since eight of these companies were still based in France at the time of their candidacy. Proof that the level of international ambition no longer necessarily goes through exile.