Advertisement|Remove ads.

Advertisement|Remove ads.
Anthropic (ANTHZZX) on Monday released its public IPO filing, which showed that more than 1,900 customers paid the Claude AI maker at least $1 million over the past year, nearly double the tally in April.
The company is targeting a public listing in November, according to The Wall Street Journal, though the timing could change. The report said Anthropic’s offering could raise $100 billion or more at a valuation of $1.5 trillion to $2 trillion.
The filing does not yet disclose Anthropic’s target valuation or the number of shares it plans to sell. Those details are expected in an updated filing after the company gauges investor feedback. Anthropic plans to list on Nasdaq under the ticker ‘ANT’.
Advertisement|Remove ads.
The filing breaks down the company’s customer base into three tiers based on revenue:
| Customer tier | Trailing 12-month revenue | Number of customers (as of Sept. 30) |
| Tier 1 | $100,000 or more | More than 11,900 |
| Tier 2 | $1 million or more | More than 1,900 |
| Tier 3 | $10 million or more | Approximately 250 |
On Stocktwits, ANTHZZX was among the platform’s top trending tickers at the time of writing, while retail sentiment around the company was in ‘extremely bullish’ territory.
Anthropic’s filing showed revenue of $16.3 billion in the first half of 2026, up from $1.16 billion in the same period a year earlier. Retail traders on Stocktwits highlighted the company’s financials.
Advertisement|Remove ads.
Revenue growth also outpaced the increase in Anthropic’s computing costs. Revenue rose 1,368% year over year in the second quarter, while compute and infrastructure costs increased 300%.
Advertisement|Remove ads.
The company reported a net loss of $80.7 billion, largely driven by a $72.4 billion accounting charge related to convertible notes. As Anthropic’s valuation increased, the value of those notes rose, creating an accounting expense.
The filing also lists $8.76 billion in “other expenses” for the first half, separate from the convertible-note charge, but does not explain what the category includes in its summary pages.
Meanwhile, spending on sales, marketing and partnerships climbed 26-fold to $3.17 billion in the first half from $120 million a year earlier.
Advertisement|Remove ads.
Goldman Sachs and Morgan Stanley will lead Anthropic’s IPO, with 20 other underwriters named in the filing. The same two banks held the lead roles in SpaceX’s offering this year.
Morgan Stanley was also named stabilization agent, responsible for managing the stock’s early trading following the offering.
The filing lists CEO Dario Amodei among the company’s largest disclosed individual shareholders, with ownership or control of 245 million shares. Daniela Amodei, Tom Brown, Jack Clark, Jared Kaplan, Sam McCandlish and Christopher Olah each own more than 200 million Class A shares, according to the filing.
Advertisement|Remove ads.
Public investors will have limited voting power under Anthropic’s proposed five-class share structure:
The structure gives the founders and the Long-Term Benefit Trust substantial influence over the company’s board, even after it goes public.
Earlier this month, Anthropic said it would allocate 5% of its shares to a public-benefit fund and could issue additional shares as the company’s valuation increased.
Advertisement|Remove ads.
The filing identifies the IPO price as the baseline for future grants. For every 50% increase in the share price above the previous milestone, the board may authorize an additional grant of up to 2% of outstanding shares. Each grant remains at the board’s discretion.
The filing also outlines risks tied to training advanced AI models, keeping pace with their capabilities and securing enough computing capacity. Anthropic said it has experienced failed or underperforming training runs that consumed resources with little or no benefit. Such runs can take weeks or months.
The company said it uses Claude to help develop future versions of the model, warning that this process could accelerate AI capabilities faster than its safety measures can keep pace, and that it might not recognize a significant acceleration until after it has occurred.
Advertisement|Remove ads.
Anthropic cited its decision to limit the initial release of its Mythos model through Project Glasswing as an example of a precaution it may take again. Delaying a release, however, could give competitors an opportunity to move ahead.
Computing capacity is another constraint. Anthropic said it has requested more capacity from cloud providers than they were willing to supply.
The filing also addresses customer complaints over changes to usage limits, including allegations that the company’s disclosures about model access were inadequate or misleading.
Advertisement|Remove ads.
Read also: Meta Launches $35B Stock Offering With Nvidia, AMD, Broadcom Backing Its AI Buildout
For updates and corrections, email newsroom[at]stocktwits[dot]com.
Comments posted here will also appear on symbol pages.