TL;DR

  • European and Israeli AI application startups attracting $30 billion in 2025 funding — 66% of US levels, up from 10% in 2016
  • US model development investment ($106 billion) dwarfs Europe ($4 billion), with infrastructure requiring trillion-dollar commitments
  • Accel partner argues European capital better deployed in applications using existing models and infrastructure rather than competing on compute

Strategic Focus on Application Layer

Philippe Botteri, partner at Accel, has outlined a clear strategic directive for European AI investment: prioritise application-layer startups over foundation model development. The firm’s 2025 Globalscape report reveals European and Israeli cloud and AI application companies are on track to raise $30 billion this year — representing 66% of US funding levels ($45 billion). This marks dramatic progress from Accel’s 2016 report, when European investment represented just 10% of US equivalents.

“The application layer is where Europe has shown it can have clear global leadership,” Botteri stated. Companies like Swedish coding platform Lovable (recent $200 million Series A) and UK-based Synthesia demonstrate European potential in building products atop existing AI models — a less capital-intensive approach than competing with US model development.

Capital Constraints Define Strategic Boundaries

The funding gap becomes pronounced in foundation model development, where US companies attract $106 billion versus Europe’s $4 billion. OpenAI’s $40 billion round and Anthropic’s $13 billion Series F illustrate the scale required for model competition. “Very few companies in the world can invest tens of billions of dollars a year,” Botteri observed, noting model development concentration in the United States.

Infrastructure requirements compound these constraints. Accel estimates $4 trillion in AI data centre capacity will be needed by 2030, with hyperscalers like Amazon, Google, Meta, and Microsoft providing most capital. The US government’s $500 billion Stargate project for OpenAI exemplifies state-level commitment unavailable in Europe. Whilst European governments pursue sovereignty through infrastructure investments — including the EU’s €200 billion AI gigafactory initiative and France’s €109 billion data centre programme — these represent fractions of transatlantic capital flows.

Looking Forward

European AI application startups are securing comparable round sizes to US peers, with German defence unicorn Helsing raising €600 million Series D and Lovable’s $200 million Series A matching American equivalents. However, questions persist regarding whether this signals sustainable leadership or an AI bubble formation. Botteri acknowledges cyclical patterns: “Whenever there is a new platform shift you get a lot of companies, some of them will stick and create lots of value and some of them will disappear.” Success depends on whether Europe’s application-layer bet produces “great leaders and champions” whilst avoiding costly infrastructure competition.

Article based on interview by Sifted

Share this article