Startups, is it still possible to get started in AI in 2025?

Artificial intelligence is experiencing unprecedented growth. Generative models, advanced automation, cross-cutting applications in industry, healthcare and education: AI now permeates the entire global economy.
At the same time, the market is structured around a few global players with significant financial resources, while environmental, regulatory, societal and European sovereignty concerns are becoming central issues.
In this rapidly changing sector, the possibility of creating and developing new AI start-ups raises questions. It was precisely this question that guided the discussions at the 2025 Start-up Festival.

AI: a dynamic but profoundly reorganised sector

The recent acceleration of AI, driven in particular by the rise of generative models, might suggest that there are numerous opportunities for creating start-ups. However, the market appears to be tense. Technologies are spreading rapidly, access to data and computing power remains highly concentrated, and significant capital has already been invested in a few powerful players.

In this context, projects wishing to emerge must not only offer technically sound solutions, but also demonstrate their relevance and ability to fit into an already very dense ecosystem. The discussions highlighted that there is still room for innovation, but that it now lies in more targeted niches, responding to the specific and deeptech needs of customers, far removed from generative AI models.

Financing an AI start-up in 2025: a delicate balance

The first round table of the day focused on the issue of financing, a critical challenge for entrepreneurs. According to the speakers, the AI sector has passed the peak of media excitement and is approaching a phase of consolidation.

Investors remain present: a large majority of funds continue to direct their investments towards projects incorporating AI. However, requirements are becoming more stringent.

The French ecosystem has recognised scientific and technical expertise, but faces significant international competition, particularly in terms of access to capital. However, financing is not the only route: several players highlighted the importance of early customers, pilot projects financed by manufacturers and strategic partnerships, which enable the use of AI to be validated before considering scaling up through fundraising.

Public initiatives, such as France 2030, play an essential role. They focus on supporting projects that will bring about profound change in key sectors such as health, the environment, education and agri-food.
The aim is not to compete with the major platforms, but to invest in areas where French and European sovereignty is essential.

AI start-ups facing societal and technological challenges

The second round table focused on the ability of AI start-ups to bring about lasting change in society. The speakers highlighted a number of key factors.

The availability of open source models has facilitated access to advanced technological building blocks, enabling small organisations to experiment quickly. This accessibility boosts entrepreneurial momentum, but also increases competition.

European regulations can be perceived as a barrier to entry. Aimed at regulating usage and strengthening the confidence of citizens and organisations, they can, on the contrary, become a competitive advantage for the creation of new start-ups.

The issue of AI's environmental impact is key. The expected increase in energy consumption (sevenfold by 2030) is encouraging the development of more energy-efficient approaches and the optimisation of resource use. The ability to innovate in this area could become a determining factor in the years to come.

What AI start-up should you launch in 2025?

Although AI is ubiquitous today, it is still possible to create a start-up in this field, but the conditions for success are more demanding. Project leaders must identify a specific (often niche) use case and develop robust and responsible solutions to meet it. It is essential to integrate issues of trust, sovereignty and sobriety into their projects.

In this rapidly changing environment, support structures such as Inria Startup Studio play a vital role in supporting creation and offering project leaders a structured framework for development.