Samsung Invests in Dutch Rival to Nvidia's AI Dominance

Samsung has committed to a Dutch startup named Euclyd as part of a 231 million dollar funding round. This investment marks a significant shift in the semiconductor industry, where major players are actively seeking alternatives to the graphics processing units that have dominated artificial intelligence workloads for years. The move highlights a growing consensus that the current infrastructure may be too energy-intensive and costly for the next phase of economic and scientific growth.

According to reporting by GN technics/ai (en-US), the funding was co-led by several major institutions including Somerset Capital Partners and EQT. The startup, founded in 2024, is developing a new architecture specifically for inference tasks. This approach aims to reduce the heavy energy demands associated with running large AI models, offering a potential solution to the rising costs faced by data centers worldwide.

Challenging the existing market monopoly

Nvidia became the world's most valuable company by repurposing gaming chips for AI training and inference. However, this near-monopoly is now facing pressure from hyperscalers and startups alike. Companies like OpenAI, Google, and Meta are developing their own processors to reduce dependency on external suppliers. Euclyd joins this group of challengers, aiming to provide a fundamentally different infrastructure that supports secure and efficient self-hosted AI systems.

Strategic value beyond financial support

For Euclyd, the partnership with Samsung offers more than just capital. As one of the largest memory manufacturers globally, Samsung brings deep expertise in engineering, supply chains, and system design. This industrial knowledge is critical for a startup that needs to transition from theoretical architecture to physical hardware. The collaboration suggests a broader industry trend where traditional chipmakers are backing innovative competitors to diversify the global technology supply chain.

Uncertain commercial timeline and risks

Despite the substantial investment, Euclyd faces significant hurdles. Its systems have not yet been proven at scale in commercial deployments, and the company does not plan to roll out physical chip systems until 2028. This long lead time means that competitors may solidify their positions in the interim. Additionally, the startup must successfully sell both hardware and intellectual property to achieve its revenue goals, a dual strategy that carries higher execution risk than focusing on a single product line.