Intel's CHIPS Act 'Rescue' Is a Loan With Handcuffs. The Government Just Became Its Co-Pilot.

On November 26, 2024, the U.S. government and Intel signed a deal the headlines had already written: a multibillion-dollar grant to bring leading-edge chipmaking home. Five days later, on December 1, the man who had staked his comeback on exactly that promise was gone — Intel's filing says he 'retired.'⁶ Two months after that, the company reported it had lost $18.8 billion for the year.⁴ The 'rescue' and the unraveling were not sequential. They were the same event, viewed from two windows.

The official story is that Washington handed Intel $8.5 billion to save American semiconductors. Almost none of that sentence survives contact with the documents. The number is wrong, the verb is wrong, and the word 'handed' is the most wrong of all.

The number everyone quotes was never signed

The $8.5 billion figure comes from a preliminary memorandum of terms in March 2024 — a non-binding handshake, not a contract.⁸ By the time the actual award was finalized in November, the direct funding had been cut to $7.865 billion: $3.94 billion for Arizona, $1.86 billion for Oregon, $500 million for New Mexico, $1.5 billion for Ohio, and $65 million for workforce programs.¹² That is roughly $640 million quietly subtracted between the press release everyone remembers and the agreement Intel signed. The gap is the first tell. Awards that get smaller between announcement and signature are not gifts being celebrated; they are commitments being renegotiated as the recipient's position weakens.

Here is the thesis, plainly: Intel's CHIPS Act money is not a bailout. It is a performance-gated loan with handcuffs, and the gating is the entire point. The Department of Commerce does not wire the cash and walk away. It releases the money in tranches, each one withheld until auditors verify physical construction progress on the ground.⁸ By early 2025, Intel had received just $2.2 billion of the $7.86 billion — $1.1 billion in the fourth quarter of 2024 and another $1.1 billion in January 2025.⁴⁷ The other 72% sits behind a turnstile that only opens when concrete gets poured.

Why a milestone gate becomes a leash

The mechanism is what turns a subsidy into supervision. A lump-sum grant gives a company capital and freedom. A milestone-gated disbursement gives it capital and a chaperone — because the only way to unlock the next tranche is to build on the government's schedule, in the government's locations, to the government's verification. The leverage runs the wrong way for Intel. Commerce can pace its money to Intel's progress; Intel cannot pace its progress to its own balance sheet. And the package came with a further constraint that bites exactly where a distressed company hurts most: dividend restrictions. Intel suspended its dividend in the fourth quarter of 2024 anyway⁴ — but the principle stands. A company fighting for survival needs maximum financial flexibility, and the terms of its lifeline narrow that flexibility precisely when it is scarcest.

Run the numbers Intel was bringing to the table. Fiscal 2024 revenue of $53.1 billion against a net loss of $18.8 billion.⁴ The third quarter alone carried $15.9 billion in impairment charges and $2.8 billion in restructuring, for a GAAP loss of $3.88 per share.⁵ This is not a healthy company accepting incentives to expand. This is a company writing down its own assets by the tens of billions, reaching for a line of credit it must earn one verified building at a time. The government isn't a savior here. It's a senior lender that gets to inspect the collateral before each draw — which makes it, functionally, a co-operator of a national-security asset it cannot afford to let fail.

The buildings that justify the money keep moving away

Nowhere is the gap between promise and pour wider than in Ohio. Intel broke ground on the Ohio One campus in September 2022, with chips promised by 2025. That target slid to 2027, then 2028, and now sits at roughly 2030 to 2031 for the first fab, with a second around 2032 — a five-to-six-year delay on a flagship project.⁷ The detail that exposes the bind: Ohio handed Intel a separate $600 million state construction grant carrying a December 31, 2028 completion deadline — a deadline Intel's own revised timeline already appears to breach.⁷ When the construction that unlocks the federal money is years behind the schedule the local money demands, the milestone gate stops being an accounting nicety. It becomes the thing standing between Intel and the cash it announced to the world.

Isn't milestone-gating just prudent stewardship?

The fair objection writes itself: this is taxpayer money, and tying it to verified progress is exactly how responsible industrial policy should work. Releasing $7.86 billion in one wire to a company losing $18.8 billion a year would be reckless; the gate protects the public from funding fabs that never open. All true. The point is not that the terms are unfair — it is that they are revealing. You attach this much oversight to a transaction only when you do not trust the recipient to execute alone, and the structure of the deal is therefore an admission about Intel's condition that the celebratory headlines obscured. There is a second, sharper objection: maybe the leash is the rescue. A purely passive grant to a company this troubled might simply disappear into losses; the milestone gate forces the capital toward the one outcome the government actually wants — buildings that make chips on American soil. That is genuinely defensible. But it also concedes the thesis. If the discipline is the value, then Washington is no longer a check-writer. It is a co-operator, setting the pace and inspecting the work, of a company it has decided is too strategic to leave to its own management.

Intel got its lifeline. It also got a co-pilot it never voted for, a schedule it cannot meet, and $5.6 billion still parked behind a turnstile that only turns when the concrete cures. The story we keep telling is that the government rescued a champion. The documents tell a colder one: a champion got measured at its weakest moment, and the price of survival was handing over the controls. The genius of the CHIPS Act was never the size of the check. It was the realization that the safest way to fund a company you don't fully trust is to never quite let go of the money — and never quite let go of the company either.