Updated
Updated · CNBC · Sep 14
Samsung Backs Euclyd in $231 Million Round as Startup Targets Nvidia GPU Alternative
Updated
Updated · CNBC · Sep 14

Samsung Backs Euclyd in $231 Million Round as Startup Targets Nvidia GPU Alternative

1 articles · Updated · CNBC · Sep 14

Summary

  • $231 million in Series A funding will help Dutch startup Euclyd develop an AI inference chip system with a non-GPU architecture, with Samsung joining Somerset Capital Partners, EQT's Scaleup Europe Fund and Innovation Industries.
  • Founded in 2024, Euclyd says its processor and memory design can cut AI data-center energy use and costs, aiming at a market seeking alternatives to Nvidia's near-monopoly in high-end AI chips.
  • Samsung's value goes beyond capital, CEO Bernardo Kastrup said, citing its memory manufacturing, engineering expertise, supply-chain reach and network as Euclyd works toward commercial scale.
  • Euclyd has not yet proven its systems in large deployments, but plans to start rolling out physical chip systems in 2028 and serve thousands of enterprise customers by 2030.
  • The raise fits a broader push by hyperscalers and startups to build in-house AI silicon, with OpenAI, Google, AWS and Meta all advancing chips for AI workloads.

Insights

Why is Samsung pouring millions into an unproven European AI architecture instead of relying on established GPU giants?
Will the massive 125kW power draw of Euclyd's new AI racks revolutionize data centers or create an impossible cooling nightmare?
Can a Dutch philosopher's untested 100x efficiency chip truly shatter Nvidia's AI monopoly by 2028?