INTC

Intel was incorporated in July 1968 by Robert Noyce and Gordon Moore, two of the engineers who had helped invent the integrated circuit at Fairchild Semiconductor. Andy Grove joined as the third employee and later ran the company through its defining growth years. The firm made its name first in memory chips, then pivoted decisively to microprocessors after Japanese competition eroded the memory business in the 1980s. The pivotal commercial event was the personal computer. When IBM chose Intel's processor for its first PC in 1981, and Microsoft's operating system to run on it, the pairing of Intel silicon and Windows software, later nicknamed Wintel, became the engine of the desktop computing boom. For roughly three decades Intel held two advantages at once. It designed the best mainstream processors, and it manufactured them in its own factories using the most advanced production process available anywhere. That combination, a chip designer that also owned the leading edge factory, is the integrated device manufacturer model, and Intel was its most successful practitioner.

The economic engine for most of Intel's history rested on that manufacturing lead. Semiconductor performance improves as transistors shrink, a trend Gordon Moore himself described, and for decades Intel shrank transistors faster and more reliably than anyone else. The company ran a cadence it called tick-tock, alternating a new manufacturing process one year with a new chip design the next, and competitors could not keep pace. Owning the factory meant Intel captured the full margin of both the design and the manufacturing, and it meant rivals who designed competing chips had to manufacture them on processes that lagged Intel's by a generation or more. This is the moat that funded decades of high profitability and made Intel one of the most valuable companies in the world.

That moat broke. Beginning in the middle of the last decade, Intel's manufacturing arm stumbled badly on the transition to its 10 nanometer process and then again on 7 nanometer, with each node delayed by years. While Intel stalled, Taiwan Semiconductor Manufacturing Company, a pure foundry that builds chips for others but designs none of its own, kept advancing on schedule. By around 2018 to 2020 the unthinkable had happened. TSMC, and to a lesser extent Samsung, had taken the manufacturing lead that Intel had held for a generation. The consequences cascaded. Advanced Micro Devices, Intel's longtime rival in PC and server processors, designs its chips and has them built at TSMC. Once TSMC's process pulled ahead, AMD could ship faster, more efficient processors than Intel and took meaningful market share in both desktops and the lucrative data center. Apple, long an Intel customer for its Mac computers, designed its own processors and left for TSMC entirely. The company that had won by owning the best factory was now losing because its factory was no longer the best.

Intel's products are organized into several groups. The Client Computing Group sells the processors that go into laptops and desktop PCs, still the largest single line and the one most associated with the Intel brand. The Data Center and AI group sells server processors and related silicon to cloud providers and enterprises. The Network and Edge group serves telecommunications and networking equipment. Together these three are reported as Intel Products. Standing apart is Intel Foundry, established as its own operating segment, which encompasses the factories, the process technology, and the contract manufacturing service that aims to build chips for outside customers the way TSMC does. Intel has also been shedding adjacent businesses to focus and to raise cash. It sold a controlling 51 percent of Altera, its programmable chip unit, to the private equity firm Silver Lake in 2025, retaining a minority stake, and it has trimmed its holding in Mobileye, the autonomous driving technology company it took public.

The foundry strategy is the center of everything. Under the previous chief executive the company committed to a plan to regain process leadership within a few years and, more ambitiously, to open its factories to external customers and become a genuine contract manufacturer competing with TSMC. This is a structural change in identity. For its entire history Intel built chips only for itself, and persuading rival chip designers to trust Intel with their most valuable designs is as much a question of business culture and customer confidence as it is of engineering. To make that credible, Intel separated Foundry into its own segment with its own profit and loss reporting, so the factories must justify their economics on their own rather than being subsidized invisibly by the product groups. The near term technical milestone is the process node Intel calls 18A, which has moved into production and underpins new chips including the Panther Lake client processor and the Clearwater Forest server processor. The longer term commercial bet rests on the node after it, called 14A, which Intel needs to anchor with a major external customer to prove the foundry model can stand on its own.

The competitive landscape is unforgiving. In manufacturing, TSMC is the benchmark, with a commanding share of the world's leading edge production and the trust of nearly every major fabless chip designer. In processors, AMD continues to press Intel in both PCs and data centers. The largest strategic threat, though, is the shift of computing spending toward artificial intelligence, where the dominant chip is the graphics processor and the dominant supplier is Nvidia. Intel arrived late and weak in AI accelerators, and a large share of the capital that once flowed into the kind of general purpose server processors Intel sells now flows into Nvidia's hardware instead. That backdrop makes the events of 2025 striking. In September 2025 Nvidia announced a 5 billion dollar investment in Intel at a set share price, paired with an agreement to co-develop products that connect Nvidia's accelerators with Intel's processors and to have Intel build certain custom chips. A rival of sorts became a partner and an investor at the same time.

Government involvement is the other defining feature of the current chapter. In August 2025 the United States government agreed to take roughly a 10 percent equity stake in Intel, funded largely by converting grant money previously awarded under the CHIPS and Science Act into common shares purchased at about 20 dollars each. The stake was structured as passive, with no board seat or governance rights, but it came with a warrant for an additional slice of the company that becomes exercisable only if Intel ceases to own a majority of its foundry business, a clause designed to keep American leading edge manufacturing intact and American owned. SoftBank separately took a roughly 2 billion dollar position. The logic is national rather than purely commercial. Policymakers concerned about reliance on Taiwan for advanced chips, and about competition with China, see a healthy Intel foundry as a strategic asset, and they backed it with public capital in a way that is rare for an American company of this size.

Leadership reflects the gravity of the situation. Lip-Bu Tan, a veteran semiconductor investor and the former chief executive of the chip design software company Cadence, became chief executive in March 2025 after a brief period of interim leadership, and he rejoined a company whose board he had previously served on. His mandate is to fix the cost structure, restore manufacturing credibility, and decide how aggressively to pursue the foundry ambition versus simply stabilizing the products business. The board itself is changing, with Craig Barratt elected as independent chair effective after the 2026 annual meeting, succeeding Frank Yeary. The presence of government, Nvidia, and SoftBank on the share register adds an unusual set of stakeholders to whom management must answer.

The risks are specific and large. The foundry turnaround requires enormous sustained capital spending on factories that may not reach competitive cost and yield, and it requires winning external customers who have every reason to stay with the proven incumbent in TSMC. If 18A and 14A do not deliver both performance and reliable manufacturing yields, the entire strategic premise weakens. The products business faces continued pressure from AMD in its core markets and from the broader migration of computing budgets toward AI hardware where Intel is not the leader. There is concentration risk in a single, expensive, multi year bet on regaining process leadership. There is execution risk in running a designer and a manufacturer that now must transact at arm's length. And the government stake, while supportive, introduces political dimensions to a commercial enterprise that markets do not always reward. Separating the foundry could unlock value or could remove the manufacturing that gives Intel its strategic distinctiveness, and that tension sits unresolved at the heart of the company.

The forward question for anyone weighing Intel is whether a company can rebuild a manufacturing lead it took years to lose, while simultaneously inventing a contract manufacturing business it never had, inside an industry where the incumbents it must displace are operating at full speed. The integrated model that once made Intel unbeatable is now the thing being tested, because the same factories that conferred the advantage became the liability when they fell behind. A successful turnaround would restore a domestic leading edge foundry and a competitive products line, an outcome that government backers, partners, and patient investors are betting on. A failed one leaves a diminished designer outsourcing its best chips to the very rival that overtook it. The capital, the partnerships, and the political will assembled around Intel are unusually large, and so is the engineering gap that all of it is meant to close. How that gap narrows or widens over the next several years is the entire story.