Competition Heats Up to Challenge Nvidia’s AI Chip Dominance

The artificial intelligence (AI) revolution has intensified competition in the semiconductor sector, as rivals seek to close the gap on Nvidia, the company that has dominated the AI chip market.

Once virtually unknown to the general public just three years ago, Nvidia now boasts the world’s highest revenues, driven by sales of its graphics processing units (GPUs) — the processors central to building AI systems such as ChatGPT and its competitors.

Although Nvidia was not the first to develop GPUs, the California-based company made them its speciality in the late 1990s, coinciding with the early days of cloud computing, giving it unmatched experience in the field.

According to Dylan Patel, head of consultancy SemiAnalysis, Nvidia is a “three-headed dragon”. The company does not merely design chips; it also provides the necessary infrastructure, networking, and software to make them work seamlessly together — the dragon’s other two heads.

“Nvidia can satisfy every level of need in the datacentre with world-class product,” said Jon Peddie of Jon Peddie Research.

Nvidia’s market share is estimated at roughly 80 percent, leaving competitors trailing far behind. Until recently, American firm AMD was considered the runner-up.

However, AMD derives most of its revenue from CPUs — processors for personal and business computers that are less powerful than GPUs. Analyst Jon Peddie notes that AMD “can’t divert resources from that golden egg.”

To reduce reliance on Nvidia, major cloud providers have developed their own processors.

• Google introduced its Tensor Processing Unit (TPU) a decade ago.
• Amazon Web Services (AWS) launched Trainium, a cloud-dedicated AI chip, in 2020.

Today, Google and Amazon account for more than 10 percent of the AI chip market and have, in some areas, surpassed AMD in performance, pricing, usability, reliability, and production capacity, according to Jordan Nanos of SemiAnalysis. Google reportedly offers its chips to third-party customers, whereas Amazon does not sell its Trainium externally.

China, the only nation rivaling the United States in the sector, is racing to catch up, but faces export restrictions on the most advanced US chips.

For Nanos, Huawei is among Nvidia’s most credible competitors, alongside Google and Amazon, and ranks ahead of AMD. Chinese tech giants Baidu and Alibaba are also developing proprietary AI processors, though these currently serve as substitutes for Nvidia GPUs.

“They can’t catch up technically for a while using in-country fabrication facilities,” said Peddie.
“But over time, with its huge and skilled workforce and subsidised investment, China will be able to produce state-of-the-art fabrication systems.”

No expert predicts Nvidia will lose its grip on the AI chip sector in the near term.

“Nvidia underpins the vast majority of AI applications today,” noted John Belton, analyst at Gabelli Funds.
“Despite their lead, they keep innovating by launching a new product every year, a pace that will be difficult for competitors to match.”

In early September, Nvidia announced its new generation chip, Rubin, set for commercial release in late 2026, with AI performance estimated at 7.5 times that of its current flagship product, Blackwell.