# Intel Government Equity Stake On **2025-08-22**, the [[Department of Commerce]] took a **9.9% equity stake** in [[Intel Corporation]] — 433.3 million shares at $20.47, roughly **$8.9 billion** — by converting money the company was already owed. It is the largest and most consequential entry on the [[Government Equity Stakes Roster]] and the deal that turned an improvisation into a policy. ## Where the money came from No new appropriation was involved. The purchase was funded out of: - **$5.7B** in unpaid [[CHIPS and Science Act]] grants Intel had been awarded but not yet received - **$3.2B** in federal awards under the Secure Enclave program for defense-grade chips In other words, the government bought Intel stock with money it had already promised Intel. ## The part that should not be skipped This is the detail that makes the deal a story rather than a transaction. Intel had **repeatedly delayed** its Ohio fabrication plants and announced factory layoffs in Oregon and Arizona through 2025. Under the CHIPS statute, failure to meet grant milestones gives the government the right to **claw the money back**. Commerce did the opposite. Per Intel's own SEC filings, the department **accelerated** the planned CHIPS payments and agreed that "to the maximum extent permissible under applicable law, the Company's obligations pursuant to [the CHIPS grant] will be considered discharged" — with a carve-out for some Defense Department obligations. The CHIPS Act was written "to incentivize investment in facilities and equipment in the United States for semiconductor fabrication." The government released Intel from the fabrication commitments and took equity instead. As Peter Harrell wrote in Lawfare, that is contrary to the spirit of the statute even if it survives the letter. ## How it is structured **Passive, on paper.** The government holds no board seat and agreed to **vote its shares in alignment with Intel's board on most matters.** This structure does real work: it is the argument against the Government Corporation Control Act problem (the government is not acquiring a corporation "to act as an agency"), and it is the reassurance to other shareholders. It is also revisable. Nothing in the arrangement prevents a future administration — or this one — from renegotiating what "passive" means. ## The unmanaged conflict Commerce now **holds Intel stock and writes semiconductor export-control policy**. Those rules determine what Intel's competitors may sell and to whom. The department is simultaneously an owner of one firm in the industry and the referee for all of them. No mitigation of this conflict has been published. ## The scoreboard argument By mid-2026 the stake was reported as worth roughly **$42 billion** — a paper gain of several multiples. This is the administration's single strongest talking point and it should be stated plainly rather than argued around: on returns alone, the Intel position has been a success so far. Two caveats belong next to it. First, the gain is unrealized. Second, a rising share price does not answer the question the CHIPS Act was asking, which was whether advanced chips get manufactured in the United States. Intel's fab timelines are the metric that statute cared about, and they slipped. ## The stake is shrinking on its own (August 2026) > [!important] The passive structure has a cost nobody priced: automatic dilution > On **2026-08-10** Intel announced a $15B common stock offering and **upsized it to $20 billion** — 210,526,315 new shares at **$95 per share**, closing 2026-08-12. **The government did not participate.** Its 433.3 million shares did not change; the denominator did. The stake fell from **9.9% to roughly 9.4–9.5%**. > > This follows arithmetically from the deal's own terms. A 9.9% holding of ~4.377B shares, against a post-offering count of ~4.588B, is 9.44%. The dilution is the structural finding, not the percentage. A passive, non-participating minority holder with no anti-dilution protection and no pro-rata participation right **loses ownership every time the company issues stock** — and a capital-intensive chipmaker in an AI buildout will issue stock repeatedly. Absent a decision to buy more, the government's position erodes automatically. Nothing was taken from it; it simply did not keep up. Note also what the offering says about the original premise. The government's shares were bought at **$20.47**; Intel sold new shares at **$95**. The taxpayer entry price was excellent — and the company demonstrably did not need to sell equity to the government to raise money on good terms. ## There is a disposition rule after all — and the clock runs out next month The roster's general claim that the program has no disposition policy holds at the **program** level. It is **not** true of the Intel deal specifically, and this page previously overstated it. Per Intel's registration filings, the government's securities **may not be transferred until the first anniversary of the closing date**, and after that may be transferred **only in broadly syndicated offerings** — including government-sponsored auctions and underwritten offerings — with escrowed shares further restricted until released. The transaction closed **2025-08-27**. Intel's prospectus supplement states the terms without ambiguity: the offer and sale of the securities "may only be made on or after **August 27, 2026**, the first anniversary of the Closing Date." > [!bug] Date correction, 2026-09-27 > An earlier version of this page gave the closing date as 2025-09-26 and the lock-up expiry as "around 2026-09-26." **Both were wrong by a month** — the error came from a secondary summary and was not checked against the filing. The Form 424B7 states "on August 27, 2025 (the 'Closing Date')" explicitly. **The lock-up lifted 2026-08-27 — a month before this correction, not a day.** There is a third restriction: the holder must give Intel **at least 30 days' notice** before selling under the registration statement. A sale cannot be sprung — it is visible in advance by contract. So there are two real constraints: a **one-year holding period** and a **channel restriction** limiting sales to broad, syndicated distributions rather than negotiated private blocks. That second one is the meaningful anti-cronyism feature in the entire program — it is much harder to hand a favored buyer a discounted block through an underwritten public offering. It exists in *this* contract; it has not been generalized into policy. > [!question] Status as of 2026-09-27: the lock-up has lifted and nothing has happened > The lock-up expired **2026-08-27**. As of 2026-09-27 — one month later — **no sale, and no announcement of an intent to sell, has been reported.** The government is now free to sell (via syndicated offerings, with 30 days' notice) and has not. > > **The plumbing was laid well in advance.** Intel filed the resale registration in **January 2026**, seven months before the lock-up lifted, registering up to **673,839,150 shares** — the 433.3M common plus the 240.5M underlying the warrant. The filing carries the standard disclaimer that registration "does not necessarily mean that the selling securityholder ... will offer to sell any of the securities." So the capability was deliberately prepared and has been deliberately unused. > > **Who decides, on what criteria, and into whose hands** remains unwritten. Every month of non-sale in a rising market is now itself a decision — one nobody has had to defend, because no policy requires anyone to explain it. ### Why holding is the more interesting choice A month of inaction while the position appreciates is not neutral. It means: - **The upside argument has replaced the exit argument.** With the stake reportedly worth several times its cost, selling would convert a growing talking point into a fixed, auditable number — and end the ability to cite a larger figure next quarter. An unrealized gain is rhetorically more useful than a realized one. - **Dilution is doing what a sale would do, without a decision.** As documented above, the position shrank from 9.9% to ~9.4–9.5% in August 2026 purely through non-participation in Intel's offering. The government is exiting *by degrees*, passively, with nobody on record as having chosen it. - **The 30-day notice cuts both ways.** It prevents a surprise dump, but it also means any eventual sale is preceded by a month in which the market knows a large seller is coming. ## Connections - The frame: [[American State Capitalism]] - The ledger: [[Government Equity Stakes Roster]] - The no-equity variant: [[US Steel Nippon Golden Share]] - The 2026 sequel: [[CHIPS Act Chipmaker Equity Round 2026]] - The statute: [[CHIPS and Science Act]] ## See Also - [[Intel Corporation]] - [[Department of Commerce]] - [[Howard Lutnick]] - [[Government Equity Stakes Roster]]