For nearly fifteen years, the economics of mobile applications was based on a relatively stable equation, namely that developing a product required time, large teams and significant capital. This rarity favored the emergence of leaders capable of dominating their category for several years. Instagram, Spotify, Duolingo and Strava have all benefited from this logic. The difficulty of building a product was itself a barrier to entry.
This equation is rapidly changing. Rocapine, a young Parisian company founded at the end of 2024 by Stanislas Marchand, Jean-Gabriel Boinot-Tramoni and Sammy Teillet, has just raised $13 million from Educapital, Daphni, Ring Capital, Center Court Capital, Athletico Ventures and Better Angle. The company already boasts $6 million in ARR, 2.5 million downloads and five successful applications in the areas of women’s health, nutrition and addiction management. Behind these results lies a more ambitious thesis using artificial intelligence to industrialize the creation of mobile applications.
So the ambition is not to build a reference application, but to build a machine capable of producing hundreds of them.
Where most publishers concentrate their resources on a limited number of products, Rocapine claims an approach based on permanent experimentation. The company plans to test up to 400 app concepts this year. When a product finds its market, it is developed, optimized and then deployed on a large scale using a largely automated technological infrastructure. One of the wallet apps reportedly reached $1 million in ARR just sixteen days after launch.
This model is more reminiscent of mobile game studios than traditional software publishers. Stanislas Marchand’s career at Voodoo is probably no stranger to this philosophy. In mobile gaming, the most successful players have long abandoned the idea of predicting the success of a product, they prefer to multiply experiments, measure results and concentrate resources on the few projects that naturally emerge from the market.
Artificial intelligence is now amplifying this logic, development costs are falling rapidly, and design cycles are shortening. Code generation, design, content or analysis tools allow small teams to launch more products in parallel. Development gradually becomes a commodity.
The model supported by Rocapine suggests a different approach from the current model, the application becomes one asset among others within a much larger portfolio. Some fail quickly, others reach critical mass, and only a few justify massive investments.
This development raises a broader question: are mobile apps becoming disposable products?
The term may seem excessive, but it reflects an emerging economic reality. If artificial intelligence makes it possible to launch hundreds of products at low cost, some applications may only have a limited lifespan. Their objective would no longer necessarily be to build a global brand over ten years but to respond quickly to an identified demand, generate revenue and then give way to a new generation of more suitable products.
The precedent of mobile gaming already illustrates this dynamic. Thousands of games are released every month. Only a minority reaches significant size. Most disappear after a few months of existence. There is no guarantee that the wellness, nutrition or personal development markets will escape this logic as creation costs continue to decline.
However, this abundance does not necessarily mean a democratization of the market. If applications become easy to produce, user acquisition becomes more complex. Value creation could then be concentrated more in the hands of distribution platforms, advertising networks and players capable of effectively exploiting data. Power would shift from developers to distribution specialists.
This is probably where the most important strategic issue lies; artificial intelligence is not only transforming the way applications are developed. It redefines the very nature of the software asset. It is no longer a question of knowing what the next big application will be but of identifying which companies will be able to build the best application factories.
