Intel priced an upsized common stock offering on 10 August, selling 210,526,315 shares at $95.00 for gross proceeds of $20bn — up from the $15bn it had proposed the same day. Net proceeds are roughly $19.7bn.
The mechanics
Underwriters hold a 30-day option on an additional 31,578,947 shares, worth about $3.0bn at the offer price. J.P. Morgan, Goldman Sachs, Morgan Stanley and Citigroup led the book, with a further ten banks in the syndicate. Closing was set for 12 August.
What Intel did not say
The release commits the money to "general corporate purposes, which may include capital expenditures and working capital". It does not mention 14A, does not name a site, and does not attach a figure to any programme. Coverage tying the raise to the next-generation node is inference; Intel has attached the proceeds to nothing.
Where the $100bn figure comes from
The widely repeated claim that the book drew roughly $100bn of demand appears in press reporting, not in Intel's release. The upsize from $15bn to $20bn is the only demand signal Intel itself disclosed.
Dilution on this scale is a choice
Issuing 210.5 million new shares — before any greenshoe — is a deliberate preference for equity over debt at a company already carrying government and partner money. It buys balance-sheet room for a multi-year capital programme without a coupon attached, and it does so while the market is willing to fund anything adjacent to AI capacity.
What money does not fix
Equity closes a funding gap; it does not close a customer gap. Proceeds buy capacity, and capacity only pays if external foundry customers commit to it. Intel disclosed no new design wins alongside the offering, which leaves it a financing event on its own terms rather than a demand signal.