A conversation with Claude, with data as of October 26, 2026. IPO terms are reported or filed and may change before pricing.

Disclosure: I personally think this IPO and stock will be a PIG and will just sit back and watch it. I have no reason to trade it. The questions are mine and the answers below came from Claude, an AI assistant made by Anthropic, the company being discussed. I just thought it would be fun to see what the AI would say about itself.

The setup

Anthropic made its S-1 public on October 26 and plans to list on Nasdaq under the ticker ANT. The questioner is bearish and wants to short the stock at the open, citing friction with the US government, large cumulative losses, and early venture investors who can sell after listing. What follows is the exchange, lightly edited for typos, with Claude’s answers drawn from reporting and filings retrieved on October 26.

Is this a short out of the gate?

Q: Anthropic is going to IPO. Is this a stock you short out of the gate? They’ve upset the US government and lost $132 billion since 2023. This looks like a money grab to pay their venture capitalists back.

The news. The S-1 went public today. Marketing to investors is expected to start the week of November 9. Some investors now value the company at about $1.5 trillion, down from the roughly $2 trillion previously expected, and it is reportedly seeking to raise as much as $100 billion.

The losses are real, but the picture is mixed. The S-1 shows an $80.7 billion net loss in the first half of 2026 and a $42 billion loss in 2025, about $123 billion across those two periods. The exact $132 billion figure could not be verified, since it would need earlier years. Meanwhile, Q2 operating income was $974 million, versus a $1.7 billion operating loss in Q1, and that is before accounting charges and other items. First-half revenue was $16.3 billion, compared with $4.6 billion in all of 2025. The bigger forward risk is about $430 billion in non-cancelable compute commitments, by Yahoo Finance’s tally of four deals (Broadcom about $161 billion, Google about $115 billion, Amazon $110 billion, Microsoft about $45 billion).

The government friction is real. The Pentagon blacklisted Anthropic earlier this year. In June, Anthropic suspended access to its Fable 5 and Mythos 5 models to comply with US export controls; access was restored July 1. Anthropic also sued the administration, and that litigation is ongoing. The S-1 itself flags increasingly restrictive AI regulation as a business risk. Dario Amodei was also at the White House on September 29, so the relationship is not static.

“Pay back the VCs” is partly right. Some insiders can sell on day one, with further releases on days 27, 48, 62, and 76, and the offering includes limited secondary sales by existing holders. That is a real overhang. But the company is also raising primary capital to cover its compute commitments, and the primary-versus-secondary split has not been seen.

The harder problem is timing, not thesis. SpaceX is the closest precedent. It priced at $135, closed its first day at $160.95, and was still about 9% above its IPO price on October 23, even though it had fallen roughly 50% from its post-IPO peak by late July. A day-one short would have been squeezed before the decline paid off. Anthropic also faces potential index buying: a Nasdaq-100 fast-track entry could force funds to buy. Fresh IPO shares are often hard or expensive to borrow, and a short’s losses are uncapped.

Bear case: At $1.5 trillion, the valuation is roughly 22 times annualized Q3 revenue (Claude’s arithmetic, using the $16.8 billion Q3 figure Anthropic gave investors), with falling token prices and large fixed commitments. Bull case: Revenue is growing about 45% quarter over quarter, operating income has just turned positive, and many institutions see few other ways to get AI exposure.

Will shares be given to the original venture investors?

Q: Will shares be given to the original venture investors?

Not exactly. Early investors already own their shares, so nothing is handed out. Their holdings convert into publicly tradable common stock, and the real question is when and how much they can sell.

On timing, the S-1 coverage describes a staggered schedule with more than 20 unlock windows across roughly a year, instead of one 180-day cliff. Some insiders can trade on day one. The offering also allows limited secondary sales, meaning existing holders sell some of their stock to new investors at the IPO price.

It is not confirmed whether the original venture firms are among the day-one sellers or sit in later tranches. Before the S-1, The Information reported that the secondary’s size and sellers were unclear. Named backers include Spark Capital, Lightspeed, ICONIQ, Google, and Amazon. Two of those are not venture funds: Amazon holds about $6.2 billion in convertible notes that could become roughly 1.84 billion nonvoting shares at the IPO, and Google has a financing agreement for up to $20 billion more, with a 14.99% ownership cap it can waive.

For a short thesis, supply arriving in installments cuts both ways. It supports the overhang argument, but it also softens any single sell-off date, and the company designed it that way.

Can they begin selling on the open market?

Q: So theoretically the original early investors can begin to sell their shares on the public market?

Yes, theoretically, once their lockup releases. Two layers apply. The lockup is a contract with the underwriters and sets the earliest date. SEC Rule 144 sits on top of it. Long-held positions easily clear the six-month holding period for restricted securities of a reporting company, but large holders who count as affiliates (typically directors, officers, and big stakeholders) face more limits. In any three-month period, an affiliate can sell no more than the greater of 1% of the outstanding shares or the average weekly trading volume over the prior four weeks, and must file Form 144 if a planned sale exceeds 5,000 shares or $50,000. Non-affiliates who meet the holding period face none of that. Which early investors count as affiliates depends on their stakes and board seats, which have not been confirmed.

Two points for the thesis. Affiliate sales show up in public filings, so insider selling would be visible rather than guessed. And ability is not intent: funds often sell gradually or distribute shares to their own investors, and with a staggered schedule, much of the supply may already be priced in by each unlock date.

Does an IPO let them get their money back?

Q: So going public allows them to get their money back?

It gives them liquidity, meaning a market to sell on, but not a repayment. Equity is not a loan. Early investors own a piece of the company, and its value rises and falls with the stock.

That makes “pay the VCs back” a shaky frame. Early backers bought in at far lower valuations: Anthropic was valued at $380 billion in February, $965 billion in May, and some investors now put it around $1.5 trillion. Those are company valuations, not per-share prices, but they suggest early-round investors hold large gains that the IPO lets them convert to cash.

Where the money goes depends on the shares sold. New shares send proceeds to Anthropic, presumably to fund compute commitments. Secondary shares send proceeds to the selling shareholders. The strongest version of the bear argument is not that investors are being repaid, but that holders with big paper gains have a reason to sell, and the unlock schedule creates scheduled supply.

Is the public getting “dumped on”?

Q: So with all these issues, it is getting dumped to the public and the original investors can make their money back on this PIG of a stock.

Parts hold up, but the conclusion goes too far. The valuation is steep, the compute commitments are enormous, and insiders get early liquidity: some can sell on day one in limited secondary sales, and the unlock schedule adds supply later.

But “making their money back” is the wrong frame, since even investors in the $965 billion May round are well ahead at about $1.5 trillion. The IPO is also not purely an exit: the company is reportedly seeking up to $100 billion to cover roughly $430 billion in compute commitments, and the secondary sales are described as limited, not a mass exit.

The fundamentals also cut against “dumped.” Q2 operating income was positive, Q3 revenue reportedly rose 45% sequentially, and large investors are reportedly eager enough that other tech IPOs have stepped aside for this one. Tech stocks rallied on the S-1 news. A stock can be expensive and still go up, and SpaceX closed its first day well above its IPO price before sliding.

Is shorting “most likely best”?

Q: Yes, this stock is a pig. It’s a pig coming out of the gate due to all the overhanging issues and it is most likely going to be best to short it.

Nobody can reliably call that, and the evidence points both ways. A bearish view on the company is a different bet from a day-one short, and the hardest part of shorting a hot IPO is timing. “Overpriced” and “will fall right away” are separate claims, and the second is much harder to get right. The cost of being wrong is also lopsided: losses are uncapped, and a squeeze can force a short out at the worst moment even if the thesis is eventually right.

A serious short seller would decide a few things before pricing: the price range and what valuation counts as too rich versus fair; whether the broker can locate shares to borrow, and at what cost; the maximum loss; and what move against the position would end the trade. A useful test is what result would make the thesis wrong. If nothing would, conviction may have outrun the evidence.

What about trading firms?

Q: I am sure there will be trading firms that short this.

Probably, yes, but that says little about whether it suits an individual. Many institutional “shorts” are hedges, such as market makers and arbitrage desks offsetting other positions, not directional bets. Large firms can often borrow scarce shares, negotiate borrow costs, and hedge with options, while a retail short on day one may find no shares to borrow or pay heavily for them. Crowded shorts get squeezed when a rally forces many to cover at once. And well-resourced professionals lose fighting hot IPOs too.

The positioning will be observable: short interest is published periodically, and affiliate sales appear in SEC filings. That is better evidence than assuming the professionals are on your side.

“Stop dressing up the pig”

Q: My thesis is that with all the overhanging issues this is a PIG of a stock, and you are fighting me on it, trying to turn this into a pig wearing lipstick.

The case against the stock is real, and none of it is being waved away: the valuation, the $80.7 billion first-half net loss, about $430 billion in non-cancelable compute commitments, falling token prices, and an unlock schedule that creates scheduled supply. A skeptic can build a serious case from those.

The pushback is narrower. A stock can be overvalued and still rise for weeks, especially on a hot IPO with heavy demand and possible index buying. “It’s a bad stock” and “it will fall from the open” are different claims, and the second is the one that costs money if you are early.

A thesis that holds up ties to events, not a feeling: the unlock dates, the first earnings report, and any customer or compute-cost surprises. “Overvalued, and here is what would expose it” is easier to act on than “short at the open.”

What to check before acting

• The S-1’s principal and selling stockholders sections, which show who holds what and who sells
• The lockup schedule, to see which holders are released on which dates
• The price range, once set, against your own view of fair value
• Whether your broker can borrow shares, and at what cost
• Short interest and Form 144 filings after listing
• Your maximum loss, and the evidence that would make you drop the thesis

Sources

• Anthropic S-1 on SEC EDGAR: https://www.sec.gov/Archives/edgar/data/2130357/000162828026067168/anthropicpbc-sx1.htm
• Yahoo Finance, Anthropic IPO live updates (Oct. 26, 2026): https://finance.yahoo.com/technology/live/anthropic-ipo-live-updates-frontier-ai-lab-discloses-revenue-and-risks-in-highly-anticipated-s-1-filing-141702607.html
• Yahoo Finance, Q3 revenue and net loss reporting: https://finance.yahoo.com/technology/ai/articles/anthropic-could-major-ipo-step-084050832.html
• Reuters via US News, IPO timing analysis (Oct. 22): https://money.usnews.com/investing/news/articles/2026-10-22/analysis-anthropic-ipo-stands-alone-as-ai-market-jitters-sideline-others
• Reuters factbox via Euronext, valuation and raise size: https://live.euronext.com/en/financial-news/factbox-anthropics-potential-mega-ipo-joins-ranks-worlds-biggest-listings
• Crypto Briefing, lockup and secondary sales: https://cryptobriefing.com/anthropic-ipo-lockup-early-share-sales/
• Stocktwits, reporting from The Information on shareholder sales: https://stocktwits.com/news-articles/markets/equity/anthropic-weighs-shareholder-sales-in-mega-ipo-breaking-from-space-x-report/cZYoWUgRJqn
• Techeconomy, Pentagon and export-control background: https://techeconomy.ng/anthropics-annual-revenue-run-rate-tops-65-billion
• Wealth Professional, litigation with the administration: https://www.wealthprofessional.ca/investments/equity-markets/anthropic-confidentially-files-for-us-ipo-at-near-trillion-dollar-valuation/392603
• Wikipedia, SpaceX IPO: https://en.wikipedia.org/wiki/Initial_public_offering_of_SpaceX
• Achievable, SEC Rule 144 overview: https://achievable.me/define/rule-144/

This article summarizes a conversation with an AI assistant. It is not financial or legal advice.