The cost of developing a new drug doubles every nine years since 1950 — that's Eroom's Law, Moore's Law inverted. Today it takes 10-15 years and billions to bring a molecule to market. AI promises to close the data loop: machine learning analyzing clinical outcomes, predicting toxicity, speeding up screens. The problem? Data is the fuel, and Europe's tank is half empty.
In the US, big pharma and startups (plus Google DeepMind, NVIDIA, OpenAI) are building vertical pipelines: proprietary data + custom models + accelerated trials. China, meanwhile, has a centralized ecosystem that bypasses privacy for speed. Europe? It has GDPR, fragmented regulation across member states, and no comparable public investment. The result: European companies — and Italian ones in particular — sit out the game or become buyers of foreign technology.
Sponsored Protocol
At Meteora Web, we see this every day with our SME clients: those who don't control their tech stack end up paying rent forever.
Same story here, at pharma scale. Spending millions on US AI platform licenses, without owning the data or the model, is signing a dependency contract. While Brussels debates the AI Act, capital flows elsewhere. Italy, with a strong pharmaceutical tradition (Lombardy, Emilia-Romagna, plus specialized SMEs), risks becoming a low-margin manufacturing hub, not an innovation driver. The problem isn't technical — it's political. We need an industrial strategy that puts data at the center, with clear sharing rules (not lock-in), and public investment in compute infrastructure and talent. Otherwise Eroom’s Law keeps hitting, and Europe stays on the sidelines.
Sponsored Protocol
For anyone developing software or managing data in pharma, the lesson is: choose stacks that give you control. If the AI platform is rented and your data sits on third-party servers, you're leasing your future. Build in-house or use open source models, but keep your data close. Just as we tell small e-commerce shops to avoid lifetime subscriptions for themes and plugins, in pharma: owning your data is the only real defense against tech dependency. The digital divide is also geographic, and we Sicilians know that well. If we don't close this gap now, in ten years we'll buy drugs designed by AI trained on other people's patients.
