> ## Content Index
> Fetch the complete content index at: https://www.dailyferment.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Anthropic's IPO page says you may get none of the shares you ask for. Oil engineers worked out in 1971 why the full fill is the one to worry about.
- URL: https://www.dailyferment.com/anthropic-ipo-retail-allocation/
- Published: 2026-11-13T13:00:00.000Z
- Updated: 2026-11-13T13:00:00.000Z
- Description: Anthropic's IPO page warns you may get none of the shares you ask for. Finnish order books and 1971 oil leases show why a full fill is the one to worry about.
- Author: Rana Bilal Zafar
- Tags: Fintech, AI, #reframe

There's one line on Anthropic's IPO page that I keep rereading. Under "Will I receive all the shares I apply for?" it says: "There is no guarantee that any investor will receive all or any of the shares they apply for in the Anthropic IPO. However, you should be ready to purchase all the shares for which you apply."

So, can you get Anthropic IPO shares? You can ask. Anthropic, PBC is selling 825 million Class A shares at an expected $70 to $78, with pricing set for 18 November 2026 and Nasdaq trading under the ticker ANT from 19 November. Its prospectus says up to 20% of the shares are expected to go to retail investors, including through Charles Schwab, Fidelity, Robinhood, SoFi and E\*TRADE. Whether you get them depends on demand, and what you get tells you something about that demand. That second part is the one I care about here.

**825M**Class A shares offered

**$70 to $78**expected price per share

**up to 20%**expected to go to retail investors

**about 6x**orders vs shares, as reported by Bloomberg

Sources: Anthropic, PBC, Form S-1/A Amendment No. 2, SEC, 10 Nov 2026; Bloomberg News via investingLive, 13 Nov 2026 (reported, not confirmed by Anthropic).

## How does the Anthropic IPO retail allocation work?

The S-1/A filed with the SEC on 10 November is plain about it. A portion of the shares will "at our request, be offered to retail investors" through named broker-dealers, acting as selling group members. Purchases through them are "subject to the terms, conditions, and requirements set by each respective entity." And the 20% comes with a hedge: "up to 20% of the shares ... though final amounts will be based on actual demand in the offering." In the risk factors the same idea appears as "a minority of shares."

Up to 20% of 825 million is 165 million shares, about $12 billion at the $74 midpoint (my arithmetic, and a ceiling, not a promise).

You don't buy in the IPO. You put in an "indication of interest" through your broker app, and the app tells you on the first day of trading what you were given. Each broker decides how. investingLive, summarising other reporting on 13 November, says allocations may be pro rata, fixed or by lottery depending on the broker, that Fidelity reportedly wants at least $25,000 in assets with the firm and Schwab $50,000, and that Fidelity warns that selling within 15 days can limit access to future deals. I haven't checked those terms with the brokers, so treat them as reported.

From request to shares
1. **10 Nov 2026** Roadshow starts; S-1/A Amendment No. 2 filed with the price range
2. **Now to 18 Nov** You put in an indication of interest through a participating broker
3. **18 Nov** Final IPO price expected to be set
4. **19 Nov** Trading expected to start on Nasdaq as ANT; your app confirms your allocation

Source: Anthropic IPO page and FAQs, read 13 Nov 2026\. Dates are Anthropic's expectations, subject to regulatory clearance.

## My first reaction: retail finally gets a real slice

Up to a fifth of a deal this size, open through the same apps people use for everything else, felt like the fair version of an IPO. The average US IPO has paid well on day one. Jay Ritter at the University of Florida keeps the long record: across 9,345 US IPOs from 1980 to 2025, the mean first-day return was 19.0%. If you can get in at the offer price, that average says you should want to.

Then the demand number came in. Bloomberg reported on 12 November, as investingLive relays it, that the offering "has drawn about six times as many orders as there are shares," and that bankers told investors existing backers would be prioritized. I haven't opened Bloomberg's piece; the headline is on Google's page one. If that ratio holds, most requests get scaled back. Which made me ask a different question: in which deals do small investors actually get everything they asked for?

![Diagram: an investor asks for 100 IPO shares and gets about 17 in a deal with orders six times the shares, but all 100 in a deal with weak demand; Finnish data show small orders earned 5.1% against an 8.7% average](https://storage.ghost.io/c/d7/b7/d7b77195-48c5-45ab-8688-51750879d1d6/content/images/2026/11/df112-fill-tells-you-card.png)

In a hot deal you get a slice; in a cold one you get everything. Illustration of pro rata rationing; Finnish figures from Keloharju, 1993.

## Where the 19% average goes missing

Guess first: across 80 Finnish IPOs from 1984 to 1989 the average first-day excess return was 8.7%. Weighted by what small orders actually received, what did small investors earn?

**5.1%**, not significantly different from zero. Large orders, weighted the same way, earned minus 4.9%. Source: Matti Keloharju, Journal of Financial Economics 34, 1993.

Matti Keloharju answered that for Finland in 1993 (Journal of Financial Economics). He had the order books for 80 Finnish IPOs from 1984 to 1989\. The plain average first-day excess return was 8.7%. But weighting each deal by how many shares an investor actually received changes it. Small orders earned 5.1%, not significantly different from zero. Large orders earned minus 4.9%. His abstract says it flatly: "average returns adjusted for the bias in allocation are lower than average unadjusted returns."

Yakov Amihud, Shmuel Hauser and Amir Kirsh found the same in Tel Aviv, where everyone was cut back by the same fraction, so the rationing is easy to see. In IPOs whose excess return turned out negative, investors got on average 61% of what they asked for. In the ones with a positive return, 23%. The 63 deals where everyone got everything they asked for had an allocation-weighted excess return of minus 6.63%.

The model behind it is Kevin Rock's, from 1986, and Amihud and his co-authors summarise it in one sentence: "uninformed investors are allocated greater quantities in overpriced IPOs and smaller quantities in underpriced IPOs." The people with better information pile into the good deals, so you get a sliver of those, and they skip the bad ones, so you get all of those.

## The cross: oil leases in the Gulf of Mexico

![An offshore oil and gas platform standing in open sea under a blue sky](https://storage.ghost.io/c/d7/b7/d7b77195-48c5-45ab-8688-51750879d1d6/content/images/2026/11/df112-offshore-platform.jpg)

An offshore oil and gas platform off northeast Brazil, a stand-in for the Gulf of Mexico leases. Photo by [Gabriel Xavier](https://unsplash.com/photos/an-oil-rig-in-the-middle-of-the-ocean-XquCLVbTYLE?ref=dailyferment.com) on [Unsplash](https://unsplash.com/photos/an-oil-rig-in-the-middle-of-the-ocean-XquCLVbTYLE?ref=dailyferment.com)

The idea has an older name, and it came from engineers, not finance professors. In 1971 three engineers at Atlantic Richfield, Capen, Clapp and Campbell, wrote about bidding for offshore oil leases. The company that wins a lease is usually the one that guessed highest about what was under the seabed, and the highest guess is the one most likely to be wrong. They wrote that "the Gulf has paid off at something less than the local credit union." Richard Thaler, reviewing it in 1988, cites a later study of 1,223 Gulf of Mexico leases issued from 1954 to 1969: only 22% were profitable.

Thaler also describes a classroom version. Boston University MBA students bid on jars of coins worth $8.00\. Their average estimate was $5.13, which was too low. The average winning bid was $10.01\. Winning meant losing about two dollars, because you only won when your guess was the wildest.

Same curse, two markets

#### Gulf of Mexico oil leases

- *How you win* your bid is the highest
- *When you win* when your estimate was the most optimistic
- *What winning tells you* others thought the tract was worth less
- *The record* 22% of 1,223 leases from 1954 to 1969 were profitable

#### IPO allocation

- *How you get shares* you ask through your broker
- *When you get them all* when demand is weak
- *What a full fill tells you* others with more information asked for less
- *The record* Tel Aviv: 61% filled in IPOs with negative returns, 23% in positive ones

Sources: Richard Thaler, Journal of Economic Perspectives, 1988 (citing Capen, Clapp and Campbell, 1971, and Mead, Moseidjord and Sorensen, 1983); Amihud, Hauser and Kirsh, NYU Stern working paper, 2001.

An IPO book runs the same logic backwards. In a lease auction you win when everyone else thought less of the tract than you did. In an oversubscribed IPO you get filled in full when everyone else wanted less of the stock than you did. Either way, getting what you asked for is information, and it arrives after you have committed. That is why the Anthropic page tells you to be ready to buy everything you request.

## What I'd take from it

None of this says the Anthropic IPO will go well or badly on 19 November. I'm not a financial adviser, and this isn't a view on whether to buy. It's about reading your own fill.

If reported demand holds near six times, retail requests will mostly come back partial, and that partial fill is the normal case, not a snub. If demand fades before 18 November and your broker suddenly fills you in full, that is the Gulf of Mexico case: you got all of it because others with more information asked for less. And because the page says to be ready to buy everything you ask for, the honest way to size a request is to ask only for the amount you'd be content to hold in that worse case.

I've written about Anthropic's prospectus twice already, on [how much of its revenue comes from two customers](https://www.dailyferment.com/anthropic-ipo-two-customers-revenue-concentration/) and on [its $895 billion of compute commitments](https://www.dailyferment.com/anthropic-895-billion-commitments/). Those were about the company. This one is about the order form, which is the part most readers will actually touch.

The claim, in one sentence

In an IPO, the number of shares you get is part of the price: a full fill usually means the people who knew more asked for less.

Established

The offering terms, the broker list and the up-to-20% retail line from the S-1/A; Ritter's 19.0% average; the Finnish and Tel Aviv allocation figures; the Gulf of Mexico lease numbers.

Inferred

That retail requests in this deal will mostly be scaled back. That rests on Bloomberg's reported six times, which I have only second hand, holding until 18 November.

Guess

Whether the winner's curse bites in this particular IPO. The evidence is about averages across many deals, and nothing here predicts 19 November.

I'll come back after 19 November, when the brokers report what retail customers actually got.

**More on Daily Ferment:** [A million people reserved a Cybertruck. About 70,000 bought one. A 1936 poll explains the gap.](https://www.dailyferment.com/cybertruck-reservations-vs-sales/) and [AI agents could end the lazy-deposit discount. UK insurance shows who pays.](https://www.dailyferment.com/ai-agents-bank-deposits-loyalty-penalty/)

## Sources

- Anthropic, [Anthropic's Initial Public Offering](https://ipo.anthropic.com/us/?ref=dailyferment.com), IPO page and FAQs, read 13 November 2026.
- Anthropic, PBC, [Form S-1/A, Amendment No. 2](https://www.sec.gov/Archives/edgar/data/2130357/000162828026072484/anthropicpbc-sx1ax2.htm?ref=dailyferment.com), SEC, filed 10 November 2026.
- Eamonn Sheridan, [Anthropic IPO six times oversubscribed: what retail investors can realistically expect](https://investinglive.com/stocks/anthropic-ipo-six-times-oversubscribed-what-retail-investors-can-realistically-expect/?ref=dailyferment.com), investingLive, 13 November 2026 (relaying Bloomberg News).
- Jay R. Ritter, [Initial Public Offerings: Updated Statistics](https://site.warrington.ufl.edu/ritter/files/IPO-Statistics.pdf?ref=dailyferment.com), University of Florida, Table 1, updated 5 November 2026.
- Matti Keloharju, [The winner's curse, legal liability, and the long-run price performance of initial public offerings in Finland](https://www.aalto.fi/sites/default/files/2018-12/winnerscurse.pdf?ref=dailyferment.com), Journal of Financial Economics 34, 1993.
- Yakov Amihud, Shmuel Hauser and Amir Kirsh, [Allocations, Adverse Selection and Cascades in IPOs: Evidence from Israel](https://archive.nyu.edu/bitstream/2451/27197/2/S-FI-01-08.pdf?ref=dailyferment.com), NYU Stern working paper, October 2001.
- Richard H. Thaler, [Anomalies: The Winner's Curse](https://kylewoodward.com/blog-data/pdfs/references/thaler-journal-of-economic-perspectives-1988A.pdf?ref=dailyferment.com), Journal of Economic Perspectives 2(1), 1988.