Euclyd Raises €200M+ for AI Inference Chip; Ex-ASML CEO Joins as Chairman

Funding Round Closes AI Inference Gap

Eindhoven-based Euclyd has secured more than €200 million ($230 million) in Series A funding to build alternative chip architectures for AI inference—a critical bottleneck as enterprises deploy foundation models at scale. The round was co-led by Samsung Ventures, Somerset Capital Partners, EQT's Scaleup Europe Fund, and Innovation Industries.

Strategic Leadership Joins Board

Peter Wennink, former president and CEO of semiconductor equipment maker ASML, has joined Euclyd as chairman. Wennink's appointment signals industry confidence in the startup's technology roadmap and brings deep expertise in advanced semiconductor manufacturing. The round represents the Scaleup Europe Fund's largest deeptech investment to date.

Challenging GPU Dominance

Euclyd's approach stands apart from the GPU-centric AI infrastructure race that has benefited Nvidia and other semiconductor giants. By building a non-GPU chip-and-memory architecture specifically optimized for foundation-model inference, the company addresses a critical need: reducing latency, cost, and power consumption for deployed AI systems. As enterprises move beyond model training into production inference workloads, specialized silicon becomes increasingly valuable.

Why It Matters

The AI infrastructure landscape is consolidating around inference as a distinct problem requiring different optimization tradeoffs than training. While Nvidia dominates training, startups like Euclyd are positioning themselves to capture the inference market—where volume, efficiency, and cost matter most. Wennink's involvement suggests the semiconductor ecosystem is taking Euclyd's architecture seriously as a potential alternative to GPU-based solutions.