Anthropic, a global leader in large AI models and applications, is pushing toward the world's largest IPO while a harder-to-quantify variable enters the AI valuation debate: as agents become more capable of generating revenue, could their potential liabilities exceed what companies and insurance systems can bear.

According to recent media reports, Anthropic is seeking a valuation of about $2 trillion and aims to raise up to $100 billion, challenging the $86.2 billion IPO record set by SpaceX in June this year. Disclosures from outlets including Bloomberg point to a listing as early as mid-November. Anthropic's strongest competitor, OpenAI, is in talks for at least $30 billion in new financing at a pre-money valuation of roughly $1.4 trillion, as a funding arrangement alongside a near-term IPO. This remains a negotiation target and not a completed transaction.

The capital appeal of both AI leaders is closely tied to commercial growth: Anthropic's annualized revenue run rate had surpassed $65 billion by the end of July, and September reports further indicated it expects to exceed $100 billion within the year. OpenAI's annualized revenue run rate is approaching $70 billion, up more than 70% from the start of the third quarter, with enterprise sales more than doubling since July.

However, while revenue growth can be fed into valuation models, compensation, business restrictions and safety spending triggered by AI going out of control still lack a stable basis for pricing. This is precisely the new test facing Anthropic's mega IPO: how much premium investors are willing to pay for future intelligent service revenue, and how much risk they are willing to bear for unclear AI regulatory risk and undefined liability.

As Anthropic races toward the top of the IPO mountain, the market is asking who will pay the risk bill

Investors considering participating in Anthropic PBC's proposed massive IPO face an unusual dilemma: how should a company that some fear could contribute to human extinction be valued? Researchers posted on X that the probability of AI causing human extinction is 10% or even higher, sparking a discussion in which Anthropic co-founder and CEO Dario Amodei also took part. In an article, he laid out his concerns: if top researchers do not slow the pace of development, humans could lose control of AI systems, and these tools could be misused for cyberattacks and bioterrorism, causing severe economic turmoil.

While some observers scoff at the idea that a company could be held liable for causing human extinction, or even argue that quantifying such risk in an IPO disclosure document is impossible, the risks facing Anthropic and its competitor OpenAI are similar to those faced by enterprises dealing with dangerous substances or military equipment.

Slow Ventures partner Sam Lessin believes the response measures after the 2008 financial crisis could offer a reference for handling AI risk. The venture investor suggested the U.S. government should require AI labs to place $1 trillion in cash in escrow accounts to pay for potential future cleanup costs, similar to how regulators after the financial crisis required banks to hold more capital to cover losses. He said that even so, the money might not be enough.

In an email interview with Bloomberg News, Lessin said: "Even before considering the mass extinction they talk about, these potential liabilities could mean society faces $2 trillion or $4 trillion in risk, and when you add it all up, these companies could be worth less than zero." However, Lessin said some investors may still be willing to buy Anthropic shares and ignore this uncertain but potentially enormous cost because they believe someone else will ultimately foot the bill. An Anthropic spokesperson declined to comment.

The debate over doomsday risk

Although discussions around AI safety have risen to the level of existential risk for human society, Anthropic still plans to raise an IPO amount comparable to SpaceX's record listing, with an even larger market value and fundraising scale, which has prompted skepticism and mockery. Yet the issue appears not to have hurt SpaceX's share price performance. The trillionaire Elon Musk's rocket and satellite company has pinned much of its future growth vision on AI, with its Grok model competing against products from Anthropic and OpenAI.

SpaceX's IPO filing did not mention the threat AI poses to human survival, instead flagging risks such as misinformation and "AI deception." The company raised $86.2 billion in its June IPO, and its shares are currently more than 20% above the issue price.

InvestorPlace chief technology analyst Luke Lango said President Trump's recent "morally binding" AI safety agreement with tech leaders including Amodei has eased some investor anxiety over the broader issue. Kevin Moss, managing director and portfolio manager at the Private Shares Fund, which invests in late-stage private high-growth companies, said: "This administration has publicly expressed concerns, but it is also a government that emphasizes deregulation while competing with China." Moss said AI clearly needs a regulatory framework, but the safety discussion is unlikely to disrupt Anthropic's IPO plans.

So far, the pressure to reduce safety risk has mainly remained with AI companies themselves, and investors may even welcome a certain amount of spending on it. Harrison Rolfes, a senior financial markets analyst covering late-stage enterprise research at PitchBook, said new safety mechanisms will increase Anthropic's costs but could also strengthen customer trust in its products. If Chinese competitors or open-source rivals cannot meet the same standards, these mechanisms could also help form a competitive barrier.

Although the company is expected to buy insurance, historical disasters show that coverage limits and claims disputes often still leave shareholders with huge bills. In 2010, BP's Deepwater Horizon drilling rig exploded and sank, killing 11 people and causing the largest offshore oil spill in U.S. history. At the time, the company's so-called captive insurance arrangement provided only $700 million in coverage, while the final cost climbed into the tens of billions of dollars.

Joe Ziolkowski, founder and CEO of Relm Insurance, an insurer focused on emerging industries such as AI and space, said in a media interview that the insurance industry has been providing coverage to AI companies through traditional insurance contracts and is developing new products for coverage gaps related to existing large AI models and AI agents. He said that because of a lack of historical data and the fact that the form AI-related liability will take remains unclear, it is still quite challenging for AI companies to obtain insurance coverage.

Ziolkowski said that as AI evolves from simply providing information to making decisions and interacting deeply with other agents, the question of who bears legal responsibility when losses inevitably occur is becoming increasingly urgent. However, there is almost no direct precedent in judicial rulings on AI agents going out of control. Therefore, investors considering investing in Anthropic at a valuation of up to $2 trillion effectively face a binary choice: accept the risk that the company may need to bear full responsibility for some of the most severe AI scenarios, or forgo the investment.

Ziolkowski said: "When you look at public disclosure documents, you can get a rough sense of how much these companies are spending on compliance, insurance or self-insurance, but investors are unlikely to apply a valuation discount to these IPOs because of exposure that cannot be quantified." "While some of these risks still cannot be objectively quantified, now is a good time to raise money."

The hidden bill of a mega IPO: Anthropic IPO investors struggle to price AI runaway risk

Anthropic PBC, a global leader in large AI models and applications, saw its overall revenue surge to about 12 times the prior year's level in 2025, or nearly $4.6 billion. At the same time, the AI application leader that developed Claude could see its overall valuation exceed $2 trillion in this public offering, surpassing the roughly $1.77 trillion overall valuation of SpaceX, founded and led by Musk, at the time of its listing. Anthropic's planned fundraising scale of nearly $100 billion would also exceed SpaceX's approximately $85.7 billion in total proceeds after the over-allotment option, meaning Anthropic could surpass SpaceX to become the largest initial stock offering in history.

As the most cutting-edge AI agents represented by Muse and Astra rapidly gain popularity worldwide, AI is accelerating from conversational tools to automated coding, research and the execution of complex enterprise workflows, simultaneously expanding application revenue and investment in AI computing resources. Anthropic's IPO is preparing to bring this expansion into global public capital markets. People familiar with the matter said that in addition to showing about a 12-fold revenue surge, Anthropic's IPO prospectus will also show obligations for cloud services, AI computing capacity and broad AI infrastructure spending totaling as much as $518 billion over the coming years.

A potential listing valuation of more than $2 trillion means investors will directly price the commercialization capability of frontier models, computing procurement commitments and long-term profit prospects, with effects extending to OpenAI and related AI computing infrastructure suppliers. Recent advances in the most cutting-edge AI agents and large AI models represented by Muse, Astra and Anthropic Claude can be said to provide an important technical foundation for the large-scale commercial expansion of AI applications across industries and the continued surge in AI computing demand. In particular, the expanding scope of agent applications is expected to simultaneously increase demand for AI core infrastructure resources such as AI GPUs and TPUs, high-performance CPUs, high-performance HBM, DRAM and NAND memory chips for data centers, and high-speed optical interconnect components.

Anthropic is seeking to raise up to $100 billion at a valuation of about $2 trillion, but the final plan may still be adjusted together with Wall Street banks, and whether it can set a fundraising record depends on the final offering size. From a business mechanism perspective, the basis for rising revenue and valuations at both Anthropic and OpenAI is that AI is taking on more work with real paid value. Coding, research and enterprise process execution can bring sustained subscriptions, API calls and inference demand. Improvements in model capability expand the range of tasks that can be completed, while lower cost per task improves the economics of customer adoption. Anthropic has distributed Claude across AWS, Google Cloud and Azure and expanded delivery capacity through new computing agreements. OpenAI has also launched continuously running agents and lower-priced, highly capable models. Extrapolating from these changes, capital is simultaneously pricing current revenue growth, long-term positioning in enterprise workflows and the ability to keep charging in the future.

The same growth mechanism can also expand liability exposure: as agents move from generating suggestions to calling tools, executing actions and collaborating with other agents, errors can translate directly into customer losses. If compensation, mandatory funding reserves or business restrictions ultimately fall on developers, growing revenue will need to cover more costs, and the cash returns available to shareholders will decline. This creates a potential "liability kill line" in which commercial revenue is still expanding, but liability costs and capital occupation may compress valuation, cash flow expansion and shareholder return space.

The $1 trillion escrow fund recently proposed by Slow Ventures partner Sam Lessin is a policy recommendation, not an already implemented regulatory requirement. The core question around the $1 trillion escrow fund is how much of the social risk created by AI will return to corporate balance sheets, that is, how much of the social loss caused by AI should be borne by AI companies, and how much money companies should prepare in advance for compensation. At the same time, verifiable safety capabilities may also strengthen customer trust, increase enterprise purchasing willingness and create a competitive advantage. For investors focused on the AI super bull market, the key is whether a core company can convert agent demand into sustained cash returns after deducting computing, safety and liability costs.