• Home
• Technology

South Korean tech giant Samsung has backed Dutch semiconductor startup Euclyd in a €200 million (approximately $230 million) funding round. The investment comes amid intensifying global venture capital deployment into viable alternatives to Nvidia’s dominant graphics processing units (GPUs) powering enterprise AI workloads.

Samsung co-led the round alongside Somerset Capital Partners, EQT-managed Scaleup Europe Fund, and Innovation Industries, with participation from several additional institutional investors. Euclyd announced that the newly raised capital will accelerate the engineering and deployment of its dedicated infrastructure built for foundational AI models.

Redefining AI Inference Architecture

Founded in 2024, Euclyd is developing a specialized chip architecture that departs fundamentally from traditional GPU design, focusing explicitly on AI inference—the computational phase where trained models execute real-time tasks for end-users:

Tackling the Memory-Power Bottleneck: By integrating the compute processor directly with memory architecture, Euclyd seeks to eliminate the severe latency and data transfer bottlenecks inherent in conventional computing.

Data Center Efficiency: The platform directly targets the ballooning power consumption, soaring hardware costs, and heavy memory requirements that currently constrain hyperscale data centers.

Diversifying Beyond Nvidia's Dominance

Nvidia has established near-monopolistic dominance over high-performance AI computing, leveraging GPU architectures originally developed for video graphics to dominate foundational model training and deployment.

With soaring compute demand straining global supply chains and tech budgets, major global technology players and investors are increasingly funding proprietary silicon and non-GPU architectures to break single-vendor reliance, drive down operational costs, and secure energy-efficient compute capacity.