Intel’s (INTC) demise started with its complacency, according to one AI chips entrepreneur.
“Intel got to a point where it became a little bit smug and thought that everything it was doing was great, even when it wasn’t,” Groq (GROQ.PVT) CEO Jonathan Ross said on Yahoo Finance’s Opening Bid.
For years, Intel’s dominance in the chip industry was unquestioned. But the rise of rivals like Nvidia (NVDA) and Advanced Micro Devices (AMD) has upended the game.
Ross acknowledged that Intel remains important to the country. “The more semiconductor companies we have that can push the bleeding edge, the better off the US is,” he said. “I do think it’s important that Intel be in a strong position.”
The chip giant is mired in a prolonged slump, and its new CEO, Lip-Bu Tan, has come under fire from Trump due to business ties with China.
But Ross isn’t ruling out a turnaround. It’ll have to eliminate its bureaucracy — as the company grew, layers of management and internal politics slowed down its innovation, Ross noted.
“You don’t get to [fix] it just by handing money out. You have to fix some of the systemic issues,” he added.
On the US semiconductor landscape, Ross said the current tariffs and incentives are a step forward for domestic manufacturers.
“It’s incredibly difficult to build chips in the United States. We’re just now building the ecosystem. Even when chips are manufactured in the United States, it’s really hard to package them in North America,” he said.
Groq’s in a unique position, as it decided to manufacture in the US since its founding in 2016, shielding it from Trump’s threatened 100% tariff on semiconductors.
Ross, who previously designed customer chips at Google (GOOG) to train AI models, founded Groq, which develops Language Processing Units (LPUs). These chips are engineered to run LLMs more efficiently than Nvidia’s GPUs, accelerating the shift from basic reasoning to groundbreaking invention.
In August 2024, Groq raised $493.5 million, boosting its valuation to an estimated $5.76 billion, according to Yahoo Finance’s private companies data tracker.
Ross said large language models are becoming less mistake-prone, expanding the areas in which they can be used.
“At some point, they’ll make so few mistakes you can use it in medicine, you can use it in law,” Ross previously said on the Opening Bid podcast.
Francisco Velasquez is a Reporter at Yahoo Finance. He can be reached on LinkedIn and X, or via email at francisco.velasquez@yahooinc.com.