In April 2025, Intel was trading at $18. Fourteen months later it hit an all-time high. The turnaround was not built by Intel alone.

In April 2025, Intel’s stock was trading at 18 dollars. The company had fired its CEO three months earlier, lost the AI chip race to Nvidia so completely that analysts had stopped including it in competitive comparisons, and was being discussed in the financial press primarily as an acquisition target or a candidate for dismemberment. Fourteen months later, on Tuesday, Intel hit an all-time high after Bloomberg reported that Apple is in preliminary talks to use Intel’s foundry to manufacture chips for its American devices. The stock climbed 14 per cent in a single session, extending a run that has now returned 175 per cent this year and more than 330 per cent since the US government took a 10 per cent stake in the company last August. The turnaround is the most dramatic in modern semiconductor history. It was not built by Intel alone.

The stake

The US government invested 8.9 billion dollars in Intel common stock in August 2025, purchasing 433 million shares at 20.47 dollars each. The money came from two sources: 5.7 billion dollars in unpaid CHIPS Act grants that were converted into equity, and 3.2 billion dollars from the Secure Enclave defence programme. The government received a 9.9 per cent stake with no board representation, no governance rights, and a commitment to vote with Intel’s board on shareholder matters. It also received a five-year warrant to purchase an additional five per cent at 20 dollars per share, exercisable only if Intel sells a majority of its foundry business. That stake is now worth approximately 36 billion dollars, a return of more than 300 per cent in nine months on an investment that nobody in Washington planned as a trade.

The government’s bet was not financial. It was strategic. Intel operates the only leading-edge semiconductor fabrication facilities on American soil. TSMC controls 64 per cent of the global foundry market and manufactures virtually all of Apple’s, Nvidia’s, and AMD’s most advanced chips, almost entirely in Taiwan. The concentration of the world’s most critical technology supply chain on an island 130 kilometres from mainland China has been described by national security officials as the single greatest vulnerability in American industrial capacity. The CHIPS Act was designed to address that vulnerability. The equity stake was designed to ensure Intel survived long enough for the CHIPS Act investments to matter.

The turnaround

Lip-Bu Tan replaced Pat Gelsinger as CEO in March 2025, three months after the board forced Gelsinger out following a contentious meeting over Intel’s failure to respond to Nvidia’s dominance in AI chips. Tan, a veteran semiconductor investor who had served on Intel’s board, inherited a company that had lost more than 60 per cent of its market value in a year. He cut 15,000 employees, restructured the foundry business into a separate subsidiary, and focused engineering resources on the 18A process node, Intel’s most advanced manufacturing technology and the first leading-edge logic process built entirely in the United States.

The results arrived faster than anyone expected. Intel reported first-quarter 2026 revenue of 13.6 billion dollars, beating Wall Street’s consensus estimate of 12.3 billion dollars by more than a billion. Earnings per share came in at 29 cents against expectations of one cent, a margin of outperformance so large that the stock jumped 24 per cent in a single day, its best session since 1987. Data centre and AI revenue grew 22 per cent year on year to 5.1 billion dollars as demand for CPUs surged alongside the shift toward agentic AI workloads that require processing beyond Nvidia’s GPUs. Foundry revenue grew 16 per cent to 5.4 billion dollars. It was the sixth consecutive quarter Intel beat expectations under Tan’s leadership.

The Apple question

Apple is in early-stage discussions with Intel and Samsung about manufacturing some of its M-series processors, according to Bloomberg, in what the industry has labelled a “Taiwan plus one” strategy. Apple has depended on TSMC exclusively since it left Samsung’s foundry in 2016. Tim Cook’s 600 billion dollar commitment to American manufacturing, announced as the American Manufacturing Program earlier this year, created both the political incentive and the strategic framework for diversifying to a US-based foundry. Intel’s 18A process, a 1.8-nanometre-class node shipping in late 2026, is the first American manufacturing technology theoretically capable of producing Apple’s chips.

The talks are preliminary. No orders have been placed. Apple has internal concerns about whether Intel’s yields and performance can match what it gets from TSMC. The most likely scenario, according to analysts, is that Apple uses Intel for lower-end M-series parts, the chips in MacBook Air and base iPad models, rather than the flagship processors in iPhone and MacBook Pro. If Apple shifted 20 per cent of its base M-series wafers to Intel, one estimate puts the foundry revenue at 630 million dollars annually. For Intel, that would be transformative. For Apple, it would be an insurance policy. For the United States government, which owns 10 per cent of the foundry that would manufacture the chips, it would be validation that the CHIPS Act investment produced exactly what it was designed to produce.

The constellation

Apple is not the only company being pulled toward Intel’s foundry by a combination of geopolitical pressure and manufacturing ambition. Intel was named as the foundry partner for Elon Musk’s Terafab, a 25 billion dollar AI chip fabrication facility in Austin that claims it will produce one terawatt of AI compute annually. Musk’s teams have contacted Applied Materials, Tokyo Electron, and Lam Research to source equipment, and Intel will contribute the 18A process node. Amazon has committed to multi-billion-dollar orders for custom AI fabric chips on 18A. Google has reportedly held discussions about foundry services. The pattern is consistent: every major American technology company is either in talks with Intel or has already signed a foundry agreement, and the common driver is not Intel’s manufacturing quality, which remains unproven at scale on the new node, but the strategic imperative to have chip production capacity inside the United States.

The US government’s escalation of chip export controls against China has reinforced that imperative. The MATCH Act, advanced by the House Foreign Affairs Committee in April, would require the Netherlands and Japan to align their chip equipment restrictions with American rules, the most aggressive semiconductor export control measure in congressional history. The legislation signals that the US intends to maintain and expand restrictions on Chinese access to advanced chip manufacturing, which in turn increases the strategic value of domestic foundry capacity. Intel is the only company positioned to provide it at scale. The government’s equity stake ensures that Intel’s survival is a matter of national policy, not just corporate performance.

The gap

Intel’s foundry market share remains below five per cent, against TSMC’s 64 per cent and Samsung’s 12 per cent. The Ohio fabrication facilities, originally scheduled for completion in 2025, have been delayed to 2030 or 2031. The foundry business lost 2.4 billion dollars in the first quarter, an improvement of 72 million from the prior quarter but still a staggering operating deficit for a business that is supposed to be Intel’s future. External foundry revenue, the money Intel earns from manufacturing other companies’ chips, was 174 million dollars in the quarter. TSMC’s quarterly foundry revenue exceeds 20 billion dollars. The distance between Intel’s ambition and its current capability is measured in orders of magnitude, not percentages.

What has changed is not Intel’s manufacturing capacity but the willingness of the most powerful institutions in the American economy to act as though that capacity will arrive. The government invested 8.9 billion dollars. Apple is in talks. Musk is building a 25 billion dollar facility around Intel’s process node. Amazon and Google are placing orders. The stock market has priced Intel not as what it is today, a company with a sub-five per cent foundry share that loses billions per quarter on manufacturing, but as what it could become if the 18A node works, the fabs come online, and the geopolitical conditions that make domestic chip production essential do not change. The 330 per cent return since the government’s investment is a bet on all three of those conditions holding. In April 2025, Intel was being measured for a coffin. In May 2026, it is being treated as a national strategic asset. The distance between those two assessments is not fourteen months of corporate improvement. It is the US government deciding that Intel cannot be allowed to fail, and the market deciding to believe it.