DigiTimes 复盘英特尔政府持股与代工战略，未披露新的订单、财务或政策增量

The economics of the foundry business have always carried a particular contradiction: semiconductor manufacturing demands enormous upfront capital and years of sustained investment, yet the companies making those investments are expected to preserve enough strategic independence to respond quickly to technology cycles and customers.

For Intel, that contradiction has become unusually explicit.
Once the standard-bearer of US chip manufacturing, Intel has spent years trying to recover manufacturing leadership lost to TSMC while simultaneously confronting Nvidia's dominance of AI accelerators and the gradual improvement of Chinese semiconductor suppliers. Its answer has increasingly depended on capital from parties whose interests extend well beyond conventional financial returns.
The most consequential was Washington's decision in August 2025 to convert previously allocated federal support into a US$8.9 billion equity investment, initially giving the US government a 9.9% stake in Intel. Intel CEO Lip-Bu Tan now insists that this does not translate into operational influence.
"They don't have any policy. They don't have any influence on the governance of the company," Tan told Nikkei , stressing that Intel remains independent and free to act in its commercial interests.
That statement is technically supported by the agreement. But the structure of the deal reveals why the boundary between support and influence is more complicated than the absence of a government board seat suggests.
Passive ownership comes with a strategic red line
Under the August 2025 agreement, Washington purchased 433.3 million Intel shares at US$20.47 apiece. The US government received no board representation, governance rights, or special access to company information and generally agreed to vote with Intel's board.
Yet the agreement also contains an unusually revealing provision: Washington received a five-year warrant allowing it to acquire up to another roughly 5% of Intel if the company ceases to own at least 51% of Intel Foundry.
That is not direct operating control. It is, however, a strong structural signal.
The US government is effectively saying that preserving Intel's majority ownership of its manufacturing operation carries strategic value. That matters because Intel had previously explored structures that could give the foundry greater separation from the product business.
The contradiction therefore lies less in Tan's governance claim than in Intel's increasingly constrained strategic environment: Washington may not tell management how to run fabs day to day, but the financing arrangement gives the government an economic mechanism tied directly to the continued ownership structure of those fabs.
Intel accepted the bargain because foundry economics leave little room for purity
Intel's willingness to accept such terms is easier to understand when viewed against the economics of its manufacturing reset.
Intel Foundry posted an operating loss of more than US$13 billion in 2024, after roughly US$7 billion in 2023, as the company absorbed the cost of new fabs, process-node development and the early ramp of Intel 18A.
The pressure has not disappeared. Intel reported a US$2.5 billion foundry operating loss in Q4 2025, followed by another US$2.4 billion loss in Q1 2026, even as 18A yields improved and EUV wafer output increased.
Tan's comparison with TSMC is therefore telling. He told Nikkei that building a competitive foundry can require a 10- to 20-year horizon, making sustained government backing particularly valuable.
The choice Intel faced was not simply independence versus state ownership. It was closer to strategic autonomy versus capital endurance.
TSMC entered 2026 with enormous scale, entrenched customer trust, and an AI-driven profit engine. Its third-quarter 2026 profit rose 69% YoY to NT$763.1 billion (approximately US$23.9 billion), while the company maintained annual capex forecast of US$60–64 billion. Intel, by contrast, has had to finance a costly manufacturing comeback while simultaneously restructuring the rest of the company.
Against that gap, ideological purity becomes expensive.
Washington is only one member of Intel's new capital coalition
The government is also not Intel's only strategic shareholder.
SoftBank invested US$2 billion in August 2025. A month later, Nvidia agreed to invest US$5 billion while entering a product partnership under which Intel will build Nvidia-custom x86 CPUs for data centers and x86 SoCs incorporating Nvidia RTX GPU chiplets.
The irony is difficult to miss: Intel's recovery is now being financed partly by the very AI ecosystem that exposed its competitive weaknesses.
Nvidia's deal does not currently require it to use Intel Foundry, but it gives Intel capital, product relevance, and a closer relationship with the company dominating AI accelerators. Washington, meanwhile, gains confidence that advanced logic manufacturing remains anchored to a US-controlled company.
Intel receives something more important than either symbolism or political approval: time.
State backing may preserve independence by limiting it
This is ultimately the contradiction at the center of Intel's latest chapter.
Government ownership inevitably raises questions over political influence, especially when industrial policy and national security have become inseparable from semiconductor strategy. Intel itself warns in regulatory filings that Washington's interests may not always align with those of other shareholders.
But the counterfactual also matters.
CSIS has argued that Intel remains effectively irreplaceable as the only US-headquartered company within reach of restoring domestic leading-edge manufacturing capability. US fabs represent roughly a tenth of global semiconductor manufacturing capacity, making Intel's survival an industrial-policy issue rather than merely a corporate one.
Tan says Intel is now about 80% through its restructuring, with advanced packaging, ASICs, and AI-related infrastructure among the next growth targets.
That makes Intel's government deal look less like an outright surrender of corporate control than a calculated compromise: accept an unusual state shareholder and an ownership-linked restriction in exchange for capital substantial enough to keep the foundry strategy alive.
The paradox is that Intel may have had to give up a degree of strategic freedom precisely to preserve the possibility of remaining an independent American chip manufacturer at all.