The next generation of Intel’s PC processors will be built by the same company that makes chips for Apple, Nvidia and Qualcomm, a concession that reshapes the balance of the semiconductor industry.
Intel’s next-generation PC processor, code-named Razor Lake, will use TSMC’s 2-nanometer N2X process, according to people familiar with the matter. The move extends Intel’s reliance on its Taiwanese rival at a moment when its own foundry business is shrinking. It also raises the pressure on AMD, which competes with Intel for the same PC market and now finds itself further down the queue for the industry’s most advanced manufacturing capacity.
The decision marks a quiet reversal of Intel’s long-held strategy. For decades, Intel manufactured its own processors, arguing that owning both design and production gave it an edge no rival could match. That argument collapsed as Intel’s process technology fell behind, and the company has progressively outsourced production of its most important products. Razor Lake, built on TSMC’s most advanced node, completes that transition for the PC line.
The shift reflects the state of Intel’s manufacturing. The company’s own process roadmap has slipped repeatedly, and its 18A node, once touted as the technology that would restore its leadership, has faced delays. Customers, including the PC makers that buy Intel’s chips, have grown impatient, and the decision to put Razor Lake on TSMC’s 2-nanometer process is an acknowledgment that Intel’s internal fabs cannot yet produce the chips it wants to sell.
TSMC’s position strengthens with every such decision. The Taiwanese company now controls roughly 70 percent of the advanced process foundry market, a share that has grown as Intel pulled back and competitors consolidated. Its 2-nanometer node, which enters high-volume production this year, is the most sought-after manufacturing capacity in the industry, with orders from Apple, Nvidia and now Intel’s PC line. The concentration of advanced manufacturing in one company, and one island, has become the defining risk of the semiconductor industry.
For AMD, the development is unwelcome. AMD designs its chips and outsources production to TSMC, competing with Intel’s PC processors using the same foundry. With Intel now also demanding TSMC’s 2-nanometer capacity, the allocation of the world’s most advanced production lines becomes a zero-sum game. AMD’s chips will still get made, but the timing and volume of its production now depend on how TSMC balances two of its largest customers, one of which is now a foundry rival.
The competitive math cuts both ways. Intel gains access to manufacturing that its own fabs cannot match, potentially closing the performance gap with AMD on the PC side. But it surrenders the differentiation that manufacturing leadership once provided, becoming, in effect, a designer reliant on the same supplier as its competitors. The profit margins of outsourced design are thinner than those of integrated manufacturing, and Intel’s financial model will be tested by the change.
The foundry implications are equally significant. Intel has been trying to build a contract manufacturing business, selling its own process technology to outside customers. That effort, already struggling against TSMC’s scale, looks harder when Intel’s own products are being made by the competitor. Customers considering Intel’s foundry services can reasonably ask why they should trust Intel’s fabs when Intel itself chooses TSMC for its flagship products.
The PC market is a large prize. Intel and AMD together supply the processors inside the vast majority of the world’s personal computers, and the next generation of machines will be built around their competing designs. Razor Lake, if it delivers the performance and efficiency Intel promises, could win back share that AMD has taken in recent years. The outcome depends less on Intel’s own factories than on TSMC’s production schedule, a dependency that would have been unthinkable a decade ago.
For the industry, the shift concentrates power further in Taiwan. The world’s most advanced chips, for phones, servers and now PCs, are increasingly made in the same fabs, and every company that designs silicon is now a customer of the same monopoly supplier. Intel’s decision to put its PC future in TSMC’s hands, analysts said, is the clearest evidence yet that the era of distributed manufacturing leadership is over, and that one company now sets the pace for everyone.
The production timeline adds urgency to the competitive picture. TSMC’s 2-nanometer capacity is being allocated among a list of customers that already includes Apple’s largest product lines and Nvidia’s data center accelerators, and adding Intel’s PC volume means the foundry must balance commitments that were negotiated years apart. Analysts said Intel’s negotiating position is weaker than it appears, because TSMC knows the PC maker has no realistic alternative for the performance it needs. That dynamic shapes pricing and allocation in ways that will show up in Intel’s margins. There is also a geopolitical dimension: U.S. officials have pressed for advanced chip manufacturing on American soil, and Intel’s decision to route its flagship product to Taiwan complicates the narrative that American companies can reshore production quickly. For a company that once symbolized U.S. semiconductor independence, the choice of foundry for Razor Lake is as much a statement about the industry’s structure as it is a product decision.
