Intel priced a $20 billion stock offering overnight, upsized from the $15 billion it announced first: 210,526,315 shares at $95 apiece, expected to close August 12 [1]. Multiply those two numbers and you get $19,999,999,925, which tells you somebody worked backward from a round $20 billion and solved for the share count. Net proceeds run to approximately $19.7 billion after fees, with J.P. Morgan, Goldman Sachs, Morgan Stanley and Citigroup leading the books and holding a 30-day option on another 31,578,947 shares, roughly $3.0 billion more at the offering price [1]. The stated use of proceeds is 'general corporate purposes, which may include, but are not limited to, capital expenditures and working capital,' which is prospectus language for fabs cost money [1].
Set the record beside that price. On August 22, 2025, the US government agreed to buy 433.3 million primary shares of Intel at $20.47 per share, a 9.9 percent stake, for $8.9 billion [2]. The money was not a new appropriation. It was $5.7 billion in CHIPS Act grants Intel had been awarded but not yet paid, plus $3.2 billion from the Secure Enclave program, converted from grants into equity [2]. Intel's own release is blunt about what the stake is not: 'The government's investment in Intel will be a passive ownership, with no Board representation or other governance or information rights.' The government even agreed to vote with Intel's board, with limited exceptions [2].
Now the arithmetic. At the $95 offering price, 433.3 million shares are worth $41.2 billion. The government paid $8.9 billion, which works out to $8.87 billion at $20.47 exactly, so Intel's rounding is honest. The difference is $74.53 per share, and $74.53 times 433.3 million is roughly $32.3 billion of unrealized gain. The position marks at 4.6 times cost in just under a year. Skeptics of the $8.9 billion figure can use the bigger denominator instead: Intel counts $11.1 billion of total federal investment once you add the $2.2 billion in CHIPS grants it had already received before the deal [2]. Divide $41.2 billion by $11.1 billion and the multiple is still 3.7, on a number that includes money the government got no shares for.
Four caveats keep this a record rather than a victory lap. First, the gain is unrealized. Nobody has sold anything, the government has given no indication it intends to, and an owner trying to unload 433.3 million shares would move the very price the gain is marked at. Second, the entry was not an investment decision in any normal sense: the dollars were already committed as grants under a 2022 statute, so the relevant comparison is not this trade versus another trade, it is the same dollars with shares attached versus without. Third, the conversion had a price of its own: the claw-back and profit-sharing provisions on the $2.2 billion already dispersed were eliminated when the agreement was signed [2]. Fourth, there is an accounting mirror here we have reported before: the escrowed-share structure behind the stake produced a $12.5 billion mark-to-market charge in Intel's July quarter, which is what happens when the taxpayer's paper gain shows up on the other side of the ledger as Intel's paper loss.
The new shares also shrink the stake. A 9.9 percent holding of 433.3 million shares implies about 4.38 billion shares outstanding before the offering. Add 210.5 million new ones and the government's slice falls to roughly 9.4 percent, and it stays near 9.4 even if the underwriters exercise their full option. The share count behind that estimate is implied rather than pulled from a filed cover page, so treat it as approximate. The direction is not in doubt: every new share sold at $95 dilutes the taxpayer's percentage while raising the market's estimate of what the remaining percentage is worth.
One more contingent claim sits in the file. The government holds a five-year warrant to buy an additional 5 percent of Intel at $20 per share, exercisable only if Intel ceases to own at least 51 percent of its foundry business [2]. At $95, that warrant is $75 per share in the money on the day the trigger ever fires, which makes it less an investment than a covenant with teeth: spinning off the fabs the deal was designed to fund would hand the government another slice of the company at a fifth of today's price.
The clean version of this story fits in one sentence. A year ago the government converted $8.9 billion of grants it already owed into Intel shares at $20.47, and this week Intel found $20 billion of buyers at $95, a price that values the taxpayer's passive, no-board-seat, slightly-diluted stake at about $41.2 billion on paper. Whatever position you took on the stake when it was signed, the arithmetic since then has only run one direction, and it is all checkable from Intel's own releases.
