Anthropic's Push to Go Public: The Timeline, the Numbers Behind the $80 Billion Loss, and the Hurdles Left

Written by Christian Bremer

On October 26, Anthropic made its S-1 registration statement public, putting one of the most closely watched AI companies on a path to the Nasdaq under the ticker "ANT." The headline number that spread fastest was a net loss of $80.7 billion for the first half of 2026. That figure deserves a closer look, because it is not the kind of loss most readers assume. This post walks through the timeline, what the filing says about cash, and the hurdles that remain before shares trade. It is a summary of public documents and press coverage, not a recommendation about the offering.

Disclosure: this post was drafted with Claude, an AI model made by Anthropic, the company discussed here. All figures come from the cited sources, and readers should verify them against the filing.

The timeline

  • May 2026: Anthropic's most recent private valuation was $965 billion, from a Series H round, according to tastytrade and Yahoo Finance coverage.

  • June 1, 2026: Anthropic submitted a confidential draft registration statement to the SEC, per tastytrade.

  • October 26, 2026: Anthropic announced that it filed a Form S-1 and intends to list Class A common stock on the Nasdaq Global Select Market as "ANT." Its announcement says the price range and number of shares have not been determined and that the registration statement has not yet become effective, so no shares can be sold yet. Goldman Sachs and Morgan Stanley are the lead underwriters, with 22 underwriters named on the cover, according to tastytrade.

  • What comes next: An amended S-1 with the share count and price range, typically filed before the roadshow, then the SEC declaring the registration effective, the roadshow, pricing, and the first trade.

  • When? Reports conflict. Bloomberg, as cited by tastytrade, said the roadshow could begin the week of November 9. Inc. reported trading is expected within about two and a half weeks of the filing, another outlet pointed to the week of November 16, and a pricing target before Thanksgiving (November 26) has also been reported. None of this is confirmed by the company. Valuation talk of $1.5 trillion to $2 trillion, and a raise that could exceed $100 billion, are press estimates, not filing terms.

The "$80 billion" in context

The filing shows a net loss of $80.7 billion for the six months ended June 30, 2026, compared with $13.0 billion a year earlier and $42.0 billion for all of 2025. Here is where it came from:

  • $72.4 billion was a change in the fair value of convertible notes, derivative liabilities, and similar instruments. The prospectus says this was primarily driven by an increase in the fair value of the convertible notes.

  • $8.8 billion was "other expense, net," which Motley Fool coverage describes as a non-cash charge tied to financing.

  • The operating loss was $693 million, down from $2.1 billion a year earlier.

Here is an analogy that may help. Suppose you promised a friend a slice of your business at yesterday's price, and the business then doubled in value. Your books now show a big loss on that promise, even though no cash left your account. The more the company is worth, the bigger the paper loss. Accountants require companies to mark these instruments to market, so a rising valuation produces a larger charge. Several outlets expect these charges to fade once the notes convert to stock around the IPO, but that depends on how the instruments are structured, and I couldn't confirm that from the filing text I could access.

So the $80.7 billion is real in an accounting sense, but it isn't a cash shortfall. It does explain why the headline looks so different from the operating picture.

What the filing and reports say about cash and operations

  • Revenue: $16.29 billion in the first half of 2026, up from $1.16 billion a year earlier, and compared with $4.59 billion for all of 2025. Second-quarter revenue was about $11.6 billion.

  • Operating results: The first-half GAAP operating loss was $693 million, while the company reported adjusted operating income of $1.62 billion and adjusted net income of $2.67 billion. Adjusted figures exclude items the company chooses, such as stock-based compensation, which was $2.2 billion in the half according to Motley Fool coverage. Reports say the second quarter was the first quarterly operating profit, at about $974 million.

  • Cash on hand: The filing shows $131.7 billion in cash, cash equivalents, and short-term investments at June 30, 2026: about $112.4 billion in cash and equivalents and $19.3 billion in short-term investments.

  • Operating cash flow: Two secondary sources report about +$3.9 billion for the first half, meaning operations generated cash. I couldn't verify that line in the portion of the filing I could read, so check the full MD&A before relying on it. Inc. also reported that Anthropic told investors it was cash-flow positive in the third quarter, a claim that is not in the S-1 itself.

  • Compute spending: Compute and infrastructure expenses were about $8.9 billion in the half, up from $2.3 billion.

The real cash question: commitments, not a shortfall

If there is a financial risk in the filing, it isn't a gap in today's cash. It is the size of what the company has agreed to pay for computing power over the next decade and beyond. The prospectus says the aggregate commitments are "more than $___," with "substantially all" related to compute and infrastructure. In the preliminary version I could read, the dollar amounts are blank, and the filing says substantially all of the remainder is expected to be incurred between 2031 and 2045.

Outside estimates fill the gap with varying confidence. One analysis cited at least $518 billion in commitments, with roughly 80% payable regardless of usage, while another outlet said that number came from an earlier confidential draft and is unconfirmed. Yahoo Finance reported separate arrangements with SpaceX's AI business of up to $120.5 billion through 2029. Until the final prospectus fills in the blanks, treat these as reported figures, not facts.

The mismatch is the point to understand. The filing says revenue is "primarily consumption-based" and much of it "is not subject to long-term contractual commitments," while compute commitments "generally extend over multiple years." Revenue that can flex, against costs that can't, is the central tension for any buyer of the stock.

The hurdles that remain

  1. Paperwork and SEC clearance. The filing is missing the offering size, share count, and price range, and the SEC must declare the registration effective.

  2. Market conditions. An IPO of this size depends on investor appetite on the day. One analysis noted that Nvidia was still about 11% below its six-month high as of October 26, and the AI trade has been volatile. The midterm elections on November 3 are another event on the calendar.

  3. Valuation. A reported $1.5 trillion to $2 trillion target compares with the $965 billion private mark from May. Investors will have to decide whether the revenue trajectory supports that jump.

  4. Governance. This is a dual-class structure with limits on public holders. Only Class A shares are sold to the public, and together they can never hold more than 49.9% of the vote. A single Class F share held by an LLC of the seven co-founders carries 50.1%, and a Long-Term Benefit Trust holds a Class T share that elects a majority of the board. The filing warns that public Class A holders "generally will not be able to affect the outcome" of shareholder votes or director elections. Supporters see this as protecting the company's mission, and critics see limited accountability for outside investors. Both views are reasonable.

  5. Revenue quality. Some customers buy Claude through Amazon, Microsoft, or Google cloud platforms, so reported revenue may overstate what the company keeps after fees, according to Motley Fool coverage. Concentration is also debated: Inc. reported that the top five customers made up nearly a quarter of 2025 revenue with none above 10% in the nine months reported, while another analysis stressed how much revenue flows through two cloud marketplaces.

  6. Lock-ups. Early reports say insider share unlocks are staggered across more than 20 windows over about a year, with some as early as days 27 to 76 after listing. That is a reported detail, so verify it against the final prospectus.

  7. Regulatory and AI-specific risks. The filing calls the regulatory and political environment "evolving and uncertain" and describes catastrophic and existential AI risks, and reports say the company believes insurance may not cover a catastrophic incident. Investors should read these sections themselves.

What this means for client conversations

  • Read the right line. A client who sees "$80 billion loss" may assume the company is running out of money. The filing doesn't support that, but it also doesn't mean the risks are small. The better questions are about the size and timing of compute commitments and how much revenue is locked in.

  • Expect volatility. Newly listed stocks can trade very differently from the offer price in the first days and months. As a general principle, a high starting valuation leaves less room for disappointment, and that is true whichever way the company performs.

  • Concentration check. Clients already own AI exposure through index funds, large-cap technology stocks, and in some cases SpaceX or Nvidia. An addition of ANT, plus possible index inclusion over time, raises the question of how much of a portfolio rides on one theme.

  • Access and allocation. Retail access to IPO shares is limited and varies by brokerage, and most investors will buy on the exchange after the first trade. Prices can differ a lot from the offer price.

  • Don't act on timing guesses. Dates are still reported ranges. Wait for the amended S-1 and the price range.

As always, this is general information, not a recommendation to buy or sell any security.

Sources

The IPO is moving quickly, so check EDGAR for an amended S-1 and refresh the timeline and valuation figures right before you republish this post.

Christian Bremer
Next

Europe's Autumn of Unrest: What's Driving It and What It Means for the Euro