Record venture investment in European AI startups masks a deeper challenge: The region’s push for technological sovereignty depends on whether governments and large corporations actually buy what those startups are selling. That tension emerged as a central theme at last month’s HumanX conference in Amsterdam, where executives argued that capital without customers leaves the continent’s AI ambitions incomplete.

What You Need to Know

European AI startups raised $23 billion in the first half of 2026, a 130% year-over-year increase that now accounts for 55% of the region’s venture funding. Yet sovereignty does not require owning every layer of the AI stack. The missing piece is demand from governments and large enterprises. Specialized chips and targeted applications, not full-stack dominance, may offer the clearest path to reducing dependence on non-European technology.

Record Funding Meets a Reality Check

Europe is awash in AI capital. A joint Crunchbase and HumanX report found that AI-focused startups raised $23 billion in the first half of 2026, up 130% from the prior year. That surge sharpens a question at the heart of the sovereign AI debate: Where should countries concentrate their resources to capture the technology’s economic value and retain control over their data? The answer, according to two executives on stage at HumanX Amsterdam, is not everywhere.

Fabrizio Del Maffeo, founder and CEO of Axelera AI, and Mehdi Ghissassi, chief product and technology officer of AI71, both argued that sovereignty comes from strategic investment in specific layers of what Nvidia’s Jensen Huang calls a “Five-layer cake”: energy, chips, infrastructure, models and applications. Competing in all layers is neither necessary nor feasible for most regions.

The Semiconductor Opportunity

For Del Maffeo, the opportunity lies in chips optimized for inference, as AI processing moves from centralized data centers to edge devices. “Artificial intelligence will expand from cloud computing, from centralized data centers, to devices closer to us in the physical world,” he said. “To enable this, you need specific chips which can run efficiently, at a lower cost, to connect these networks.”

Five-year-old Axelera AI already sells two generations of inference chips to roughly 600 customers and plans to expand into decentralized cloud computing. Del Maffeo’s message was direct: Buying these chips domestically strengthens the local AI ecosystem more than importing them.

  • Axelera AI: Focuses on efficient inference chips for edge devices and decentralized cloud computing.
  • AI71: Partners with large organizations and government agencies to drive AI transformation in the UAE.
  • The Five-layer cake: Jensen Huang’s framework describing energy, chips, infrastructure, models and applications as distinct layers of the AI stack.

Government Mandates as a Demand Engine

Abu Dhabi-based AI71 offers a different model. The UAE has the largest compute per capita in the world and has mandated that every government agency set up agentic processes for citizens within three months. Over two years, the goal is to have AI agents handle half of all citizen-government interactions, according to Ghissassi.

He said competing at the model layer does not make sense when energy is abundant and cheap. Instead, governments can drive demand by becoming anchor customers. That purchasing power, combined with targeted hardware and application development, creates a self-reinforcing cycle that funding alone cannot produce.

Why This Matters

Europe’s sovereign AI push will succeed or fail based on whether it can convert record startup funding into actual adoption. Without large-scale buyers from government and industry, the region risks building a vibrant investment ecosystem that produces little domestic influence over AI’s trajectory. The examples from Axelera AI and AI71 show that focusing on specialized chips and government procurement mandates can create real leverage. For European policymakers and business leaders, the immediate question is no longer how to raise money but how to spend it on homegrown solutions.