Why Intel’s U.S. Government Stake Changes the Chip Foundry Calculus

A chipmaker’s most valuable asset is not a factory. It is the belief, held by customers who spend years designing products around a manufacturing process, that the factory will still be there and still be competitive when their chips are ready. Intel has spent the last several years struggling to sell that belief to anyone outside its own walls.
The U.S. government’s decision in August 2025 to convert roughly $8.9 billion in CHIPS Act grants into about a 10% equity stake in Intel changes what the company is selling. Washington is no longer just a subsidizer writing checks and setting conditions. It is the largest shareholder in a company that wants to be America’s answer to TSMC, and every potential foundry customer now has to price that fact into its decisions.

What the Deal Actually Does
The structure matters more than the headline number. The government received about 433 million shares at roughly $20.47 each, funded by unpaid CHIPS Act grants and money earmarked for the Secure Enclave program, on top of grants Intel had already collected. It also received a five-year warrant for an additional 5% at $20 a share, exercisable only if Intel stops owning at least 51% of its foundry business.
That warrant is the quiet center of the agreement. It does not ban a foundry spinoff, but it makes one expensive and politically awkward. For years, one of the standing questions around Intel was whether the manufacturing arm would eventually be separated, sold or merged to stop the bleeding. The warrant tilts the answer toward keeping design and manufacturing under one roof.
Why Customers Care About Ownership
Consider the position of a fabless chip designer deciding where to build a future product. Moving to a new manufacturing process means committing design teams, software tools and years of validation to a specific foundry’s rules. If that foundry changes hands, cancels a node or loses its financing mid-project, the designer pays for it in delays that can run to a full product cycle.
Intel has given customers reason for caution. Its Ohio campus has been pushed back repeatedly, and the company has said openly that its next-generation 14A process depends on landing outside customers. A foundry that tells the market it may not build its next node without enough orders is asking customers to take a leap and promising only that it might catch them.
Government ownership addresses part of that fear, though not all of it. A shareholder that has put billions of dollars into Intel and tied part of its return to the foundry staying intact has every incentive to keep the manufacturing business alive. The implicit message to customers is that the lights are unlikely to go out on this one, because too much political capital is now attached to it.
Then there is the other side. Some customers will find the arrangement reassuring, and others will find it unsettling. A company that sells chips globally may not love tying a core supplier to Washington’s priorities, especially in an administration comfortable with tariffs, export controls and public pressure on individual executives. Intel itself has acknowledged that government ownership could complicate dealings with foreign governments and customers.

The Pressure on TSMC and Samsung
For TSMC, Intel’s balance sheet was never the real threat. The threat is political. American customers and American lawmakers have grown uncomfortable with the fact that the most advanced chips in the world are made on one island, and the stake gives them a domestic option to point to when they argue for diversification.
TSMC has responded by expanding in Arizona, which blunts the argument but does not erase it. A fab on American soil still belongs to a Taiwanese company, still relies on a parent company’s research and still sits inside a corporate structure Washington does not influence. Intel can claim something TSMC cannot: it is American-owned, American-run and now partly owned by the American public.
What Still Has to Go Right
None of this guarantees orders. Ownership changes the risk calculation for customers, but it does not change the physics. A foundry wins business by delivering yields, performance and delivery schedules that match or beat the alternatives, and customers who have worked with TSMC for years know exactly what that standard looks like.
The government is also a passive shareholder in name, having agreed to vote with Intel’s board with limited exceptions. That promise is easy to keep when the stock is rising and the company is hitting milestones. It is harder to hold when a politically sensitive decision arrives, such as a plant closure, a layoff round or a deal with a customer Washington dislikes.

The 14A node is the test that will matter most. Intel needs enough outside commitments to justify building it, and the government stake is meant to make those commitments easier to sign. If a major designer puts a flagship product on that process, the equity deal will look like the move that tipped the market. If none does, Washington will own a tenth of a company that spent public money to build capacity nobody wanted to use.



