Samsung Finds a Successor to Nvidia, US Dominance Shaken
Jakarta, CNBC Indonesia – South Korean technology giant Samsung is leading a funding round worth US$231 million (approximately Rp 3.4 trillion) for a Dutch start-up that is developing a new AI chip. The chip is designed with an architecture different from Nvidia’s and is claimed to be far more energy-efficient and cheaper.
If successful, this technology has the potential to replace the dominance of Nvidia’s graphics chips, which have so far been the backbone of the entire global artificial intelligence industry, according to a CNBC International report on Tuesday (15/9/2026).
The start-up, named Euclyd, was founded in 2024 and is led by Bernardo Kastrup. The company develops specialised chip systems for AI application needs, known as the inference stage. Unlike Nvidia’s chips, which were originally made for video gaming and later adapted for artificial intelligence, Euclyd’s chips were designed from the outset specifically for AI.
The architecture is also different, so the chips are expected to work faster with far lower electricity costs.
Nvidia is now the most valuable company in the world thanks to graphics chips that have become the industry standard for artificial intelligence. However, as demand for these chips has soared, two major problems have become increasingly pressing:
Soaring electricity costs. Nvidia chips require enormous amounts of power to operate, causing data centre management costs to balloon.
An architecture not built for AI. Nvidia’s chips were fundamentally designed for gaming graphics, so they are not fully efficient for AI workloads.
These weaknesses are what Euclyd aims to solve with Samsung’s full backing.
“Artificial intelligence is becoming the foundation of economic growth and scientific discovery. But its potential will remain limited as long as the foundation underpinning it is not fundamentally changed,” said Bernardo Kastrup, Chief Executive of Euclyd.
Euclyd’s chip architecture combines the processor and memory in a single, mutually optimised design. The aim is that data no longer needs to travel far from memory to processor, drastically saving time and electrical power.
Samsung is not only providing funding, but is also contributing memory manufacturing expertise, systems knowledge, and an extensive supply chain network.
“Samsung can help us with more than just money. They are one of the largest memory manufacturers in the world. They understand systems thoroughly, they understand supply chains, and they have a very extensive network,” Kastrup added.
A Host of Companies Poised to Replace Nvidia
Samsung’s move turns out to be in line with what other major technology companies are doing. Dependence on Nvidia’s chips has left many parties feeling insecure, which is why nearly every major technology company is now trying to develop its own replacement chips:
OpenAI has just announced its first chip, named Jalapeño, claimed to offer the highest speed and efficiency in the industry.
Google continues to refine its range of specialised artificial intelligence chips, known as TPUs.
Amazon Web Services and Meta are also developing their own chips for their data centre needs.
This means the AI chip market, so far dominated entirely by Nvidia, is now beginning to shift. Samsung and Euclyd are arriving with a new architecture expected to lower costs while reducing the world’s dependence on a single company.
Euclyd’s chip systems are planned to be rolled out in phases from 2028. By 2030, the company aims to serve thousands of enterprise customers. However, the technology has not yet been tested operationally on a large scale, so it remains to be proven whether its performance can genuinely match or even surpass Nvidia’s proven chips, which are already installed worldwide.
Source: jawawa.id (https://jawawa.id/newsitem/samsung-finds-a-successor-to-nvidia-us-dominance-shaken-1789473262)
Originally published by CNBC Indonesia (https://www.cnbcindonesia.com/tech/20260915140947-37-768076/samsung-sudah-ketemu-pengganti-nvidia-dominasi-as-digoyang), translated from Indonesian