Monster IPOs Upending the World of Underwriting

From: J. Ari Pandes and Jay R. Ritter

To: IPO underwriting watchers 

Date: November 6, 2026 

Re: Monster IPOs Upending the World of Underwriting

Anthropic is poised to go public in two weeks in what might be the world’s largest initial public offering ever, with OpenAI expected to follow. When it does, one thing to watch is whether Anthropic copies the recent SpaceX IPO playbook to lower commissions and better control allocations of shares. Doing so would threaten the lucrative business model of the banks that are the biggest IPO underwriters.

Banks make money in two ways when they underwrite IPOs. The obvious one is the commission paid by the issuer, known as the gross spread. With almost all moderate-size IPOs in the United States, that number is 7 percent. SpaceX negotiated 0.666 percent, the lowest a US IPO has ever had on record, amounting to $500 million (all figures US) on its $75-billion offering. Almost all IPOs also carry a 15-percent overallotment option, allowing underwriters to sell 15 percent more shares at the same commission if it is exercised. SpaceX negotiated zero commission on those shares, lowering the average gross spread to an even smaller 0.58 percent, demonstrating SpaceX’s negotiating power.

This gross spread revenue, after subtracting expenses, is normally split among the underwriters roughly in proportion to the number of shares each underwrote. There were 22 underwriters for the SpaceX IPO. Goldman Sachs and Morgan Stanley each underwrote 20 percent of the shares, so if the net commission revenue was $400 million after expenses, each would have netted $80 million. At the other end, 12 banks, including RBC Capital Markets, each underwrote about 0.4 percent of the shares and would have received roughly $1.6 million each. Not the kind of dollars you would expect from the biggest IPO ever.

But the spread is only half the story. The second and less obvious source of profits comes from commissions paid by institutional investors on other trades. Most IPOs are oversubscribed. Normally the “lead-left” underwriter – the bank listed on the left side of the top row on the cover page of the IPO prospectus – gets to decide, in consultation with the issuing company, how many shares each institutional investor receives when demand exceeds supply. More than 90 percent of the shares are usually allocated by the lead-left underwriter, no matter what percentage it has underwritten. The other banks collect fees and provide research coverage, but receive few if any shares to allocate to clients. (BMO Capital Markets, RBC Capital Markets and TD Securities are all on Anthropic’s prospectus that was released last week.)

Underwriters tend to favour their most profitable clients, many of which are hedge funds. That encourages those clients to pay high commissions on other, unrelated trades – payments known as “soft dollars” – so that they could become such profitable clients. The more underpriced IPO shares that there are for the lead-left underwriter to hand out, the more that some investors will be willing to overpay on these other trading commissions. And unlike the gross spread revenue, none of it is shared with the other banks.

This indirect compensation does not appear in any prospectus. It surfaces later as sales and trading revenue of the banks, which makes its scale easy to miss.

In the past five years, total gross spread revenue from operating company IPOs in the United States and Canada, not including capital pool companies and special-purpose acquisition companies (SPACs), was a little more than $20 billion before expenses. Assuming that costs eat up $5 billion of that $20 billion, that puts net commissions at around $15 billion.

Money left on the table, defined as the dollar value of first-day paper profits for investors who bought underpriced shares at the offer price, amounted to more than $50 billion. If hedge funds and others competing for favourable IPO allocations paid soft dollars equal to 30 percent of the money left on the table, underwriters would collect another $15 billion as indirect compensation – a number roughly equal to their net commissions.

Until this year, this arrangement seemed to be a permanent fixture of IPOs. The SpaceX IPO upended that.

With most IPOs, about 40 percent of shares are allocated to hedge funds. With SpaceX, hedge funds were squeezed down to about 10 percent. Most of the shares went to buy-and-hold investors, including institutions that don’t pay soft dollars.

If underwriters can’t reward their hedge fund clients with underpriced IPO shares, those clients will be unwilling to overpay on other trades. Should Anthropic and OpenAI insist on most of their shares being allocated to buy-and-hold investors rather than hedge funds and are aggressive in negotiating a low gross spread, the profitability of the IPO business for major underwriters will be under attack.

Ari Pandes is an associate professor of finance and an associate dean at the University of Calgary’s Haskayne School of Business and a research fellow at the C.D. Howe Institute.

Jay R. Ritter is the Joseph B. Cordell Eminent Scholar Chair Emeritus and the director of The IPO Initiative at the University of Florida’s Warrington College of Business.

To send a comment or leave feedback, email us at blog@cdhowe.org.

The views expressed here are those of the authors. The C.D. Howe Institute does not take corporate positions on policy matters.

A version of this Memo first appeared in The Globe and Mail.