The original logic was easy enough to grasp. At the time the CHIPS Act passed, East Asia accounted for roughly 75% of global chip production, Brookings found in December 2022, and the U.S. had a thin domestic base in leading-edge logic chips. Intel was the one American company still trying to build that kind of capacity at home.

The money matched the scale of the problem. The Commerce Department said in March 2024 that it had reached a preliminary agreement with Intel for up to $8.5 billion in direct funding, up to $11 billion in loans, and access to a potential 25% investment tax credit, Commerce reported. In November 2024, it finalized up to $7.865 billion in direct funding for projects in Arizona, New Mexico, Ohio, and Oregon, with disbursements tied to milestones, Commerce said.

Intel was also the biggest bet in the program. Commerce said the award would help support nearly $90 billion in expected U.S. investment by the end of the decade, alongside about 10,000 manufacturing jobs and 20,000 construction jobs, Commerce reported.

The Trump announcement in August 2025 changed the politics of the deal, and also its shape. Instead of leaving Intel with the usual grant-and-loan structure, the administration said the government would turn the promised subsidies into equity, taking 433 million shares at $20.47 apiece, below the market close of $24.80 cited by the Competitive Enterprise Institute, CEI reported.

The exact legal mechanics of that shift are not fully laid out in the public material available here. What is clearly documented is the outcome: the government ended up with a large equity position, at a discount that came at existing shareholders’ expense.

Intel’s own filing language makes the bigger point. In an SEC filing, Intel said the government’s stake “reduces the voting and other governance rights of stockholders” and may limit future transactions that could benefit them, Cato reported. The filing also said the government may vote “as it wishes” under certain conditions, including if Intel tried to change or end its relationship with the U.S. government, Cato reported.

That is where the story changes. A grant can help fund a factory. A shareholder can shape the boardroom.

The deal appears to go beyond ordinary ownership in another way. Cato reported that the government also received an option to buy an additional 5% of Intel at $20 per share if the company sells part of its foundry business, and said Wall Street Journal reporting described that as a way to discourage Intel from fully exiting manufacturing, Cato reported.

That kind of clause matters because it ties public money to a particular corporate outcome. It does not simply support Intel’s balance sheet. It gives Washington a tool that can shape how Intel thinks about restructuring, partnerships, and whether to keep parts of its business at all.

The precedent is what has set off so much of the reaction around the deal. After the Intel announcement, Commerce Secretary Howard Lutnick floated similar equity arrangements for defense firms including Lockheed Martin, Palantir, and Boeing, Cato reported. Cato also noted that the Pentagon already holds a 15% stake in rare-earth company MP Materials, CEI reported.

That does not make Intel unique. It makes it suggestive.

The line between subsidy and ownership sounds abstract until a boardroom has to live with it. Subsidies help fund projects. Equity with voting rights changes who gets nervous when management considers a spinoff, a sale, or a retreat from a politically favored line of business.

Cato’s concern is blunt: with the government as a large shareholder, Intel will face constant pressure to align corporate decisions with whichever political party is in power, Cato reported. That pressure does not have to be explicit to work. Companies know how to read the room, especially when the room includes the U.S. government.

That is why the comparison to the auto bailouts keeps coming up. Cato argued that the 2008-2009 rescues cost taxpayers 40% more than the Obama administration claimed, or $14 billion once interest was counted, and that $61 billion could have gone to more productive uses such as retraining autoworkers, Cato reported. Intel is a different case, obviously. National security is not the same thing as the auto industry in a recession. Still, the basic worry is familiar: once the government owns a stake, exit becomes harder than entry.

There is also a market signal here. Cato said Intel’s stock had already spiked 20% since news of the equity deal leaked in early August, Cato reported. That move can be read two ways, as relief that Washington is backing the company or as evidence that investors think the public balance sheet is now part of Intel’s capital structure. Those are not the same thing.

For all the noise around the deal, some of the most important facts remain murky in the public record. There is no single authoritative accounting of the full package, including grants, loans, tax credits, and equity. The exit terms for the government stake are not clearly established in the sources here, and it is not clear when or how Washington could unwind its position.

The same goes for the project milestones that were supposed to justify the original CHIPS funding. Commerce said the money would be released based on Intel’s completion of those milestones, Commerce reported, but the public reporting available here does not show whether they have all been met. And while Commerce projected that the program would help America produce roughly 20% of the world’s leading-edge chips by the end of the decade, Commerce said in March 2024, that remains a projection, not a settled result.

That gap matters. The CHIPS Act was sold as a security and jobs program, not an ownership program. If the government stake ends up improving Intel’s competitiveness, the deal may look shrewd in hindsight. If it mostly locks Washington into managing a private company by other means, the bill will be harder to square with the original pitch.

For now, the Intel deal reads less like a finished policy and more like a trial run. The next question is not just whether Intel can build the fabs. It is whether the federal government has just found a way to become a shareholder without admitting that is what it is doing.