Staying on top in the AI chip race won’t be easy for market darling Nvidia (NVDA) and its founder Jensen Huang, reminds one fellow tech billionaire.

“The fact that Jensen doesn’t even make his own chips — that everyone has Taiwan Semiconductor (TSM) available — is all the more credit to him at the design level, the way they have done things is pretty fantastic.

Added Gates, “But so far he has stayed way ahead and Microsoft is an incredible customer and is always wanting to get as many of his chips as they possibly can. At the same time other big tech companies are working on their own AI chips.”

Competition for Nvidia is sprouting up, as Gates correctly points out.

With giants like Amazon (AMZN) announcing an $8 billion partnership with Anthropic to enter the AI chip space and Google (GOOG) dropping a supercomputer with an AI chip called Willow, it’s evident Big Tech companies want in on the action.

Further, Broadcom (AVGO) and Marvell (MRVL) have released advanced custom chips.

At the same time, the long-term demand for Nvidia’s powerful AI chips is being questioned arguably for the first time.

DeepSeek is a Chinese company that has bursted onto the tech landscape seemingly out of nowhere.

It surprised markets and those trading hot AI names like Nvidia and AMD (AMD) after unveiling RI, its AI model that gave a ChatGPT-esque performance at a cheaper price tag. RI costs a reported $5.6 million to build a base model, compared to the hundreds of millions of dollars incurred at US-based companies such as OpenAI and Anthropic.