Samsung joins $231M Euclid AI chip round

Euclid’s chief executive said Samsung was a co-investor alongside Somerset Capital Partners, Innovation Industries and the ScaleUp Europe Fund. The startup, founded in 2024, is building an inference chip around a different architecture from Nvidia’s GPUs and says its silicon could sharply reduce power demand in AI data centers. It plans to begin with physical systems in 2028 and reach thousands of customers by 2030, a long runway that underlines both the technical difficulty and the potential commercial prize.

For Samsung, the bet is strategic as well as financial. As one of the world’s largest memory-chip makers, it has a direct interest in where AI infrastructure spending flows, and could offer manufacturing know-how, supply-chain access and systems expertise if Euclid gains traction. That makes the move less about a single startup and more about Samsung positioning itself for a market in which AI chips, memory, packaging and server integration are increasingly intertwined.

The backdrop is a broader industry shift. OpenAI last month unveiled its own AI chip, while Google, Amazon Web Services and Meta are all accelerating custom silicon programs. Those moves do not necessarily dislodge Nvidia in the near term, but they do narrow the field of buyers who must rely on its GPUs for every workload. Nvidia’s own shares have been volatile as investors weigh sustained AI demand against the risk that customers will increasingly design around it.

That tension is visible in market positioning. Nvidia’s stock has slipped back toward its 50-day moving average after a strong run earlier in the year, while technical readings such as RSI and MACD have cooled from earlier overbought levels. Competitor AMD has also been volatile even after a sharp rally in recent months, reflecting a market that still believes AI spending is real but is less willing to pay any price for it. Sentiment data from Adalytica.com shows extreme fear around both Nvidia and the broader AI theme, despite awareness remaining very high, a sign that investors are still chasing the story but with much less conviction.

For investors, the key question is not whether Nvidia loses its lead overnight, but whether the next phase of AI spending broadens into a more fragmented ecosystem. If inference workloads become more cost-sensitive, startups like Euclid could gain relevance, and Samsung’s involvement would give it optionality across memory, foundry and systems. If the technology fails to scale, the investment is still a relatively small call on a market that remains captive to Nvidia’s software and hardware stack.

The next catalyst is execution. Euclid must prove its architecture can move beyond promise, secure customers and withstand the scale, reliability and software hurdles that have buried many AI-chip challengers. Until then, Samsung’s move is best read as a strategic hedge against a future where Nvidia remains central, but no longer uncontested.