The underwriting discount: what IPO banks actually get paid
Every IPO prospectus cover page lists a fee few readers notice. Holtec Nuclear's Nasdaq debut today is a real deal to walk through how it works.
Holtec Nuclear Corporation lists on the Nasdaq Global Select Market and Nasdaq Texas today, September 18, under ticker HNUC — the roadshow that launched September 8 offering 50 million Class A shares in a $15–$18 range, roughly $825 million at the midpoint against a targeted $10.2 billion valuation. Nine banks worked the deal: J.P. Morgan, Guggenheim Securities, Goldman Sachs, Citigroup and BofA Securities as joint book-running managers, with Morgan Stanley, Cantor, BMO Capital Markets and Oppenheimer & Co. rounding out the syndicate. Every one of those banks gets paid out of a single line on the prospectus cover page that almost nobody outside the deal room reads closely: the underwriting discount.
The mechanic is a markup, not a commission
An underwriting discount isn't a fee the company writes a check for. It's a spread. The underwriters agree to buy the shares from Holtec at one price and sell them to investors at a higher one — the IPO price everyone sees quoted. The difference between those two numbers, expressed as a percentage of the offering, is the "gross spread," and it's how the entire syndicate gets compensated for pricing the deal, building the order book, and taking on the risk of holding shares between purchase and resale, even if that risk window is measured in hours.
That structure matters because it means the discount never shows up as a separate charge working against a specific holder — it's baked into the capital the company actually receives. If Holtec's spread runs at a typical rate, the company nets something noticeably less than the $825 million headline figure per share sold; the gap is the underwriters' pay, deducted before the proceeds ever hit the balance sheet.
Why 7% became the industry's answer
Academic work on this — most notably Jay Ritter's long-running dataset on IPO underwriting economics — has documented a striking pattern in the US market for years: a large share of mid-sized IPOs cluster their gross spread at almost exactly 7%, regardless of company, sector, or year. Researchers have called this "the 7% solution" because the clustering is too tight to be pure coincidence in a competitive market — for deals in the few-hundred- million-dollar range, 7% functions almost as a norm the industry defaults to rather than a number each bank negotiates from scratch.
That norm bends hard at scale, though. Very large offerings routinely pay well below 7%, because the dollar amount at even a much smaller percentage still funds the syndicate's work. Alibaba's 2014 IPO — one of the largest in history — priced its spread near 1.2%, and that still worked out to over $260 million for the banks involved. It's the same logic that will apply whenever Anthropic's targeted IPO in the pipeline prices at the scale currently being reported: a raise approaching $100 billion would command a gross spread far below the norm that applies to a deal Holtec's size, simply because the absolute dollars involved make a smaller cut enormously profitable on its own.
Splitting the spread across a nine-bank syndicate
A deal with as many underwriters as Holtec's needs its own internal formula for dividing the pay, and the standard convention breaks the gross spread into three pieces: a management fee for structuring the deal, an underwriting fee compensating the syndicate for the capital risk, and a selling concession — typically the largest single piece, often around 60% of the total spread — paid out to whichever bank actually placed each block of shares with an investor. Lead-left banks, the ones named first and running the books (J.P. Morgan and Guggenheim in Holtec's case, per the prospectus structure), collect a disproportionate share of the management and underwriting fees on top of their own selling-concession allocation, which is why the "bookrunner" title on a filing is worth more than a listing position — it determines who gets paid first and most, not just whose name appears highest.
Why the fee is easy to miss and worth reading anyway
None of this changes what a share of HNUC is worth once it starts trading — the discount is a transaction cost borne once, at issuance, not an ongoing drag on the stock. But it explains something that trips up a lot of first- time prospectus readers: the difference between the amount a headline says a company "raised" and the amount that actually lands on its balance sheet. It's the same category of easy-to-skip mechanical detail as the over-allotment option — Holtec's own greenshoe, explained in detail here — sitting a few lines away on the same cover page, doing a completely different job: the greenshoe manages price stability after the deal, while the discount is what pays for getting the deal done at all.
The exact percentage Holtec agreed to isn't headlined in the press coverage of the raise, and prospectus cover pages report it as a dollar figure per share alongside the percentage, buried well below the price range that actually makes the news. That's true of nearly every IPO that crosses Zorro's upcoming pipeline — the fee is disclosed, public, and almost never discussed, right up until the deal is the one you're trying to underwrite yourself. Once HNUC starts trading, it joins the listings tracked on Zorro's recent-IPO page, where the underwriting economics behind it stop being visible and the only thing left to watch is the price.