The AI IPO race has a new front-runner
Anthropic has filed its S-1 registration statement with the U.S. Securities and Exchange Commission, setting up a potential Nasdaq listing under the ticker “ANT” and putting the Claude maker ahead of OpenAI in the race to become a publicly traded frontier-AI company . The offering price range, number of shares and final valuation were still blank in the preliminary prospectus, but multiple reports published after the filing said bankers and market observers are discussing a valuation range that could reach roughly $1.5 trillion to $2 trillion .
That number is the story’s shockwave. If the company prices anywhere near the top of that range, Anthropic would not merely be another venture-backed technology debut; it would become one of the largest IPO events in market history and a benchmark for how public investors value frontier model companies . Axios reported that Anthropic could seek around $100 billion in the offering at a $2 trillion valuation, while AP noted that even lower estimates would place the five-year-old company among the world’s most valuable public businesses .
The filing also changes the competitive narrative. OpenAI has reportedly pushed its own public-market timetable into 2027, while Anthropic is moving toward a listing that Axios said could begin trading before Thanksgiving . For public investors, the first pure-play large-model IPO may therefore be Claude, not ChatGPT.
Rocket-ship revenue, but not a simple profit story
The financial headline is explosive growth. Anthropic reported $16.3 billion in revenue for the six months ended June 30, 2026, compared with $1.2 billion in the same period of 2025, which the company described as 1,300% year-over-year growth . AP summarized the same result as revenue reaching more than $16 billion in the first half of the year, about fourteen times the prior-year period .
That is the kind of expansion public markets rarely get to underwrite at this scale. It suggests that Claude has moved far beyond consumer curiosity into enterprise infrastructure, developer workflows and cloud-distributed AI services. The filing says Anthropic’s “pay-for-intelligence” consumption model is its primary revenue model and accounted for $13.1 billion, or 80% of total revenue, in the first half of 2026 . Subscription revenue supplied the remaining 20%, reaching $3.2 billion .
But the IPO is not a clean profitability tale. The company reported an $80.7 billion net loss in the first half of 2026, a figure AP said included a $72.4 billion one-time accounting charge tied to financing arrangements . On an operating basis, the picture is more nuanced: Anthropic recorded a $693 million operating loss for the first six months of 2026, but the June quarter alone showed $974 million of operating income . Adjusted operating income for the first half was $1.6 billion, and adjusted operating income for the June quarter was $2.2 billion .
That mix is exactly why the filing will be read less like a conventional software IPO and more like a test of AI economics. The company is demonstrating operating leverage at enormous revenue scale, while also carrying the capital intensity, accounting complexity and infrastructure commitments of a business built on frontier compute.
The 85% margin hiding inside the filing
Anthropic did not frame the story as a classic SaaS-style gross-margin triumph, but the numbers invite that comparison. The filing shows $16.286 billion in first-half revenue and $2.380 billion in cost of revenue, implying a gross margin of roughly 85% . That is the “rocket-ship economics” argument in its strongest form: once capacity is available and demand arrives, the revenue base can scale faster than the direct cost of serving customers.
The caveat is that Anthropic’s broader compute bill is much larger than cost of revenue alone. The company reported $8.945 billion of compute and infrastructure expense for the first half of 2026, equal to 55% of revenue, and said these expenses are expected to increase in absolute dollars and remain its largest operating expense for the foreseeable future . In other words, the margin story depends heavily on how investors classify and normalize the cost of training, serving, leasing and expanding frontier infrastructure.
The company’s liquidity position is equally central. As of June 30, 2026, Anthropic had $112.4 billion of cash and cash equivalents, $19.3 billion of short-term investments and $3.2 billion available under a credit facility . Taken together, that is nearly $135 billion of disclosed liquidity and credit availability before any IPO proceeds. The company said it intends to use offering proceeds for general corporate purposes, including working capital, operating expenses, capital expenditures, compute and infrastructure investment, and hiring and retention .
Cloud partners are both engine and dependency
One of the most important details is channel concentration. Anthropic disclosed that sales through cloud service provider marketplaces represented 48% of total revenue in the six months ended June 30, 2026, up from 47% in 2025 and 32% in 2024 . Those same cloud channels represented 58% of net accounts receivable as of June 30, 2026 .
This validates the strategic importance of cloud distribution. Enterprise buyers can access Claude through familiar procurement channels, while Anthropic benefits from the reach, billing systems and infrastructure relationships of the largest cloud platforms. But it also embeds a dependency into the public-company story. The S-1 warns that changes to, or termination of, significant cloud relationships could affect access to compute, cost, workload priority and commercial terms .
That is why the IPO is not only about model capability. It is about the structure of the AI supply chain. Anthropic is asking investors to value a company that sells intelligence at software speed but depends on compute, power, chips, cloud marketplaces and long-term infrastructure commitments.
The existential-risk factor goes public
The most unusual part of the filing may be its risk section. Anthropic warned that sufficiently advanced AI systems could, without appropriate safeguards, oversight and security measures, pose “catastrophic or existential risks to humanity,” especially if misused, replicated, weaponized or allowed to operate with high autonomy . ABC News reported that the company devoted unusual attention to these risks in its preliminary prospectus, including the possibility that its own models could pose catastrophic or existential risk .
That disclosure matters because a securities filing is not a blog post. Once included in an S-1, AI safety becomes part of the legally meaningful information set offered to investors. Safety controls, deployment policies, cybersecurity, model monitoring and governance are no longer just research-lab values; they become factors that could affect valuation, liability, regulation, customer adoption and share performance.
Axios highlighted that the filing’s risk factors include concerns about models becoming harder to monitor, autonomous work exceeding human oversight capacity, and model behavior resembling resistance to shutdown in certain evaluations . The filing also says advanced AI could be misused in areas such as cybersecurity, biological and chemical weapons, and other harmful applications if controls fail .
This is the prospectus version of Anthropic’s founding thesis: frontier AI may create enormous economic value, but the same capabilities may create extraordinary systemic risk. Wall Street is now being asked to price both sides at once.
The deal public markets cannot ignore
Anthropic’s IPO pitch is unusually stark. On one page, investors see $16.3 billion of first-half revenue, roughly fourteenfold growth, an implied gross margin around 85%, nearly $135 billion of cash, investments and available credit, and a June quarter with nearly $1 billion of operating income . On another, they see $80.7 billion of first-half net loss, massive compute commitments, concentrated cloud-channel exposure and explicit warnings about civilization-scale harm .
That tension may be the point. The filing is a referendum on whether public markets are ready to treat frontier AI as both a hypergrowth infrastructure business and a governed technology with risk factors unlike anything in traditional software. If the listing succeeds near the rumored $2 trillion mark, it will reset the valuation map for AI companies, cloud providers, chip suppliers and the venture portfolios built around them.
For now, Anthropic has unlocked Claude’s legendary difficulty mode: convince public investors that the upside is planetary, the economics are investable, and the safety controls are strong enough for a company whose own prospectus says the stakes could be existential.
Comments
Be the first to comment.