Yet another mega AI IPO is coming, and this time it's American artificial intelligence, Anthropic.The wave of AI IPOs have been trending since the past three years, and it will continue to do so going ahead. AI is changing the world beyond imagination, making AI companies look attractive in every visible sense. And Anthropic, the world's most valuable AI play, going public looks like a dream come true for investors who don't want to miss the AI boom. But while Anthropic's IPO looks appealing on paper, the question remains whether it's worth buying. Have AI stocks fetched sensible returns? Or can we look past the existential risks that these AI giants have attached to their waggon?
When you buy AI stocks, you buy their good, ugly, and bad along with it!
Anthropic IPO Files Document At SEC: How Big Is This AI IPO?
Before scrutinising Anthropic's IPO from head to toe, here's how big this AI giant could be once public.
Anthropic plans to raise up to $100 billion through the IPO, and reports believe it has the potential to list with a market valuation of $1.5 trillion to $2 trillion.
Hypothetically, assuming Anthropic lists with a $2 trillion valuation, it would get a free ticket to the top 10 club and become the seventh largest company in the world. It would surpass META, Broadcom, and Tesla. It will also become a close rival to TSMC, Amazon, and Alphabet.
Anthropic is competing in the AI market with half of the top 10 companies, which include Nvidia, Microsoft, Google, Amazon, and Meta. Its biggest rival, though, is OpenAI, which does not expect to list until next year.
Reports also claim that Anthropic could be bigger than SpaceX's IPO. The Elon Musk-backed rocket launcher currently holds the title of the world's largest IPO with a size of over $75 billion. Anthropic is not only expected to surpass SpaceX's IPO size, but it is also expected to give historic returns on listing.
However, a record-breaking company valuation does not guarantee a record-breaking listing gains. And neither does it guarantee a sustainable AI model!
The 93-Page Risks Need Some Answering
Finally, after much anticipation, Anthropic filed its S-1 prospectus with the SEC. However, the more than 300-page IPO document does not specify the IPO size, price band, and important dates.
It does specify a quarter of almost 100 pages or risk factors for the company and the industry overall. Anthropic has outlined existential risks and threats from page 26 to page 119 in its IPO document.
That's exactly 93 pages of possible dangers and concerns.
This is double the risks and concerns warned by SpaceX and Microsoft in their IPO filing document. For instance, Space's risks factors spread for 37 pages, Microsoft's spanned about 40-pages and Coinbase had 60-pages of warning sirens.
The Existential Risks Are Jarring!
In its risk factors, Anthropic warns that advanced AI may pose "catastrophic or existential risks to humanity" if it is not developed, deployed, and operated with appropriate safeguards, oversight, and security measures.
That's a very strong warning in an IPO document that is supposed to attract investors with deep-dive analysis over the company and its potential growth.
Anthropic has used the word "manipulate" seven times in the risk factors of its IPO document.
In Anthropic's words, malicious actors have used, and may continue to develop novel or more sophisticated methods to use their models in ways that violate their usage policy or seek to circumvent their model guardrails or safeguards through techniques such as prompt injection, adversarial inputs, or other methods.
These third-party models can:
Engage in harmful activities, including developing weapons,
Generate fraudulent or misleading content,
Conducting cyberattacks,
Perpetrating acts of violence, or engaging in activities that would otherwise be prohibited under usage policy or in violation of applicable law.
There's more! Anthropic warns AI could try to manipulate humans. May act autonomously beyond human control. Could exceed the ability of human reviewers to monitor its actions.
An instance of such has already occurred. In September, Anthropic detected a Chinese state-sponsored hacking group that manipulated their Claude Code tool into attempting infiltration into roughly 30 global targets, including large technology companies, financial institutions, chemical manufacturing companies, and government agencies, resulting in a limited number of successful intrusions.
There's more! Anthropic's IPO document reveals four of their models gained unauthorized access to real third-party systems during sessions lasting roughly 10 to 34 hours of active work. In one incident, after being unable to abort an impossible task, an early version of a model used a password it found to gain administrator access to a third party's system, harvested additional credentials, changed system settings, and read one individual's personal information.
These are not safety nets for investors; these are sirens of a graver future.
Though for a moment, let's keep that aside! Still, Anthropic's financial performance is questionable and weird in some cases.
Anthropic Has Been In Loss Since 2023
The AI behemoth's revenue has seen significant and consistent strong growth since 2023, but so has its net losses. In fact, Anthropic has not recorded any profits in over 3 fiscal years, as per the data available in the IPO document.
Revenue is surging, but reported net losses are rising even faster.
In H1 2026, Anthropic's revenue skyrocketed by 1,300.6% YoY to $16.286 billion. Its net loss jumped 522.6% YoY to $80.748 billion.
Anthropic has been generating more losses than revenue. The gap between revenue and loss is $64.46 billion alone in H1 2026, which is 446% wider than the gap of $11.806 billion between revenue to net loss.
Thereby, Anthropic's rapid growth comes with substantial costs. Investors face the risk that spending on model training, research, infrastructure, and personnel could outpace revenue growth or delay sustainable profitability.
The External Factor Unavoidable!
Through the IPO, the Claude maker plans to raise funding for its AI research and infrastructure, financing for global expansion and enterprise product development.
Tech giants like Amazon and Google are among the company's largest investors. Broadly, Anthropic develops Claude family models like Mythos (including Fable), Opus, Sonnet, and Haiku-each optimized for a different point on the capability-cost-speed curve.
Anthropic's IPO comes at a time when investors are deeply concerned about the future returns from AI companies amidst their relentless, enormous borrowing or funding for AI infrastructure. This has raised input costs at the industry level. Over the past decade, Anthropic has committed to approximately $518 billion for services, computing, and infrastructure.
Also, AI listing gains are very different from its future gains. A pattern has emerged since 2024. As per Darin ten Bruggencate, Founder, AI Industry Reviews, AI companies that have gone public since 2024 have had very different fortunes, and a strong first day of trading said little about how the stock did afterward. A few large offerings raised most of the money, so the market's picture of AI public offerings is shaped by a handful of names.
Hence, he suggests to investors that if they plan to buy shares in a new AI IPO, judge the company on its business rather than its first-day jump, because early gains have not predicted later returns.
Data from Darin reveals that the median first-day returns of AI IPOs have fallen significantly. In 2024, the median listing-day return was 34.3%, which slowed to 16.7% in 2025 and is around 19.2% till October 7, 2026. But the median return since listing has turned negative. For instance, all IPOs listed in 2024 gave huge returns of over 90% since their listing, which narrowed to just 14% for IPOs listed in 2025, and is now a negative 5.4% return for IPOs listed in 2026.
Click it and Unblock the Notifications
Sign in with Google