European AI‑chip funding surge
European AI‑chip startups are drawing record investor funding as firms like Fractile and others seek to position themselves as alternatives to dominant suppliers. One rival has publicly said it's seeking at least $100 million to scale, even as observers warn that manufacturing, software compatibility and customer qualification remain hard problems. (cnbc.com)
European artificial intelligence chip startups are chasing much bigger funding rounds as investors look for alternatives to Nvidia in data centers. (cnbc.com) Dutch startup Euclyd told CNBC on April 17 that it is discussing a round of at least 100 million euros, or about $118 million. The company is backed by former ASML chief executive Peter Wennink, and says it is building chips for AI inference, the stage when a trained model answers users’ prompts. (cnbc.com) U.K. startup Fractile is pitching a different route into the same market. Its website says it is designing systems for “frontier model inference” and claims its hardware could run advanced models up to 25 times faster at one-tenth the cost. (fractile.ai) Inference is the part of AI that runs after training, when a chatbot writes a reply or an image model generates a picture. That has become the next target for chip startups because Nvidia’s graphics processors were adapted from gaming and still dominate both training and serving models. (cnbc.com) Investors are already putting real money behind that bet. CNBC reported that AI chip startups raised $8.3 billion globally in 2026, citing Dealroom, after European AI startups raised a record $21.6 billion in 2025. (cnbc.com 1) (cnbc.com 2) The pitch is not just about speed. Euclyd says its multi-chiplet design, which splits a processor into smaller connected pieces, could cut power use sharply; Fractile says cloud providers want hardware that lowers cost and reduces dependence on one supplier. (cnbc.com) (fractile.ai) Europe still has hard limits. CNBC reported that the region’s venture market is smaller than the United States, and its semiconductor manufacturing base is thin, leaving startups dependent on outside foundries even if they design chips at home. (cnbc.com) Winning customers is another hurdle. Euclyd said it is working with Samsung’s semiconductor ecosystem partners on initial test chips and is in talks with four prospective clients, with possible deliveries in 2027 or 2028. (nltimes.nl) Software is the other gatekeeper. Nvidia’s lead is not just the chip itself but the tools around it, and CNBC reported in 2024 that the company held roughly 70% to 95% of the AI accelerator market as rivals tried to match its hardware and developer ecosystem. (cnbc.com) That leaves Europe’s new chip hopefuls in a long race: raise enough cash, tape out working silicon, prove it to customers, and fit into the software stacks companies already use. The funding is arriving faster than before; the harder test is whether any of these startups can ship at scale. (cnbc.com)
