The Chipmaker’s Dilemma: Intel’s Fight for Survival in a Market Dominated by Nvidia

The gleaming new fabrication plants Intel is building in Chandler, Arizona, will eventually be visible if you drive south of Phoenix on a hot afternoon. They rise out of the desert like glass and concrete cathedrals. Clean rooms humming with machinery worth more than most businesses are traversed by workers dressed in white bunny suits. This place has received close to $30 billion. Prior to his departure, Pat Gelsinger referred to it as the company’s Agincourt. He was correct about the stakes, but the metaphor was a bit dramatic. The future of Intel will be decided here, or it will be quietly laid to rest.

The tale of Intel’s journey is the kind of corporate catastrophe that has been taught in case studies for decades. The company missed the wave twice in a twenty-year period. First, it was smartphones; Intel ignored the iPhone era until it was too late because it was so preoccupied with PC and data center chips. Artificial intelligence followed. The modest GPU was transformed into the engine of the AI economy by Nvidia, a company formerly primarily known for video game graphics cards. The future had already been consumed by Jensen Huang’s empire by the time Intel realized what was going on.

Without hesitation, the numbers convey the story. The current value of Nvidia exceeds $4 trillion. Intel is valued at about $100 billion. There isn’t a gap there. That solar system is distinct. And in a moment that seemed almost unreal, Nvidia declared in September of last year that it would pay $5 billion for a roughly 4% share in Intel. Some referred to this partnership as a collaboration, others as a lifeline, and some discreetly as something more akin to charity. That headline had a certain melancholy quality. The business that established Silicon Valley took an investment from the business that took its place.

Since assuming the role of CEO in 2025, Lip-Bu Tan has been remarkably open. On his first earnings call, he told analysts, “This is not a quick fix,” and the room practically let out a sigh. Instead of purchasing AI technology, he is attempting to develop it internally, something Intel hasn’t really done in years. For more than ten years, the company bought AI startups, including Movidius, Nervana, and Habana Labs, in the hopes that one of them would figure out the solution. Mobileye was effective. The others didn’t. Tan seems to realize that the old playbook is no longer relevant.

The more difficult question is whether Intel can continue to compete in a market where Nvidia sells entire ecosystems rather than chips. These days, the product consists of cables, software compilers, and complete data center stacks. In the meantime, Google and Amazon are stealthily developing their own AI chips to put pressure on Intel in a different way. In August of last year, the U.S. government acquired a 10% stake in a private chipmaker, turning it into something akin to a strategic national asset. This is another awkward situation. Depending on who you ask, it may help or hinder.

It’s difficult not to feel a peculiar mixture of admiration and discomfort as you watch this happen. Modern computing was created by Intel. It demonstrated to the world the capabilities of a microprocessor. It is now operating 20% leaner and relying on a single manufacturing process, 18A, in the hopes that Chandler’s desert plants will be able to accomplish what billions of dollars in acquisitions were unable to. It might be effective. Perhaps Tan’s unique strategy identifies the clients and workloads that everyone else overlooked.

However, sentiment is not rewarded in the chip industry. Execution is rewarded. Additionally, Intel’s clock is running faster than the company would like to acknowledge.