Intel has finished most of the internal overhaul it began after years of lost ground, but chief executive Lip-Bu Tan is warning that the harder part — rebuilding a competitive contract chipmaking business — will take years. In an interview with Nikkei in Tokyo earlier this month, published this weekend, Tan said the company's restructuring was roughly "80% done."
That figure covers the organizational reset, not the technology recovery. Since taking the top job in March 2025, Tan has replaced much of Intel's senior leadership. Among the recruits is Seok-Hee Lee, the former chief executive of South Korea's SK Hynix, whose memory-chip experience speaks directly to one of the areas where Intel has struggled to keep pace with rivals.
The shake-up follows a costly expansion under Tan's predecessor, Pat Gelsinger, who stepped down in December 2024. Gelsinger committed billions of dollars to building a contract manufacturing business, and those outlays helped drive a net loss of 16.6 billion dollars in the third quarter of 2024. In the years before that, Intel surrendered its lead in cutting-edge production to Taiwan Semiconductor Manufacturing Co. and lost ground in artificial intelligence processors to Nvidia.
Tan was careful not to put a timetable on the foundry effort. He pointed to TSMC's own history, noting that it needed 10 to 20 years to become the industry's dominant contract manufacturer. Building fabs and developing each new production node demands multiyear capital spending, and Intel must prove to outside customers that it can deliver competitive processes reliably and at volume — something an internal reorganization alone cannot demonstrate.
Government backing, Tan argued, is part of what makes that long horizon possible. "From a foundry's point of view, Intel had to look at the longer-term horizon, and it's always very important to have very strong government involvement supporting," he said. Last year the U.S. government converted previously allocated federal funding into an 8.9 billion dollar investment, taking a 9.9 percent stake in Intel without board representation; that stake has since fallen to about 8.2 percent following a subsequent share offering. Private backing has followed as well, with Nvidia buying 5 billion dollars of Intel shares and SoftBank buying 2 billion dollars' worth.
For nearer-term growth, Tan pointed to advanced chip packaging for AI systems and to application-specific integrated circuits — custom chips designed for a particular job. Both areas let Intel sell into the AI buildout without first having to match TSMC node for node, and both are natural extensions of a foundry business that needs anchor customers while its leading-edge processes mature. The business could become a growth engine "if we do it right," Tan said, but "it takes time."
The distinction matters for enterprise buyers watching Intel's road map. An 80 percent complete restructuring suggests the leadership and cost decisions are largely settled. It does not mean the foundry question is settled. Whether Intel becomes a credible second source for advanced manufacturing will be decided by yields, capacity and customer wins over the next several process generations — a contest Tan is now openly measuring in years, not quarters.