Anthropic’s $2 trillion IPO is too extreme to be sustainable

AI giant Anthropic has repeatedly delayed its IPO since the company first filed with the US Securities Exchange Commission in July.

The timing remains uncertain, but the latest plan is to launch a $100 billion offering just before the US Thanksgiving holidays, when Wall Street dealmaking slows dramatically until the New Year.

Capital market metrics suggest the IPO’s timing is as flawed as its $2 trillion valuation is extreme.

The accelerating rise in US Treasury bond yields amid widening wars in the Middle East, the worst energy supply shock in history with no imminent prospect of a peace deal with Iran, as well as unsettled government debt markets in Europe and Japan, hardly represent a supportive economic backdrop.

Furthermore, a surge in Oracle credit default swaps and periodic plunges in SoftBank shares, driven by investor anxiety over its $40 billion stake in OpenAI, bode ill for Anthropic’s whopping valuation. With operating losses since 2024, the firm’s net profit in 2025 was only $4.6 billion, despite exponential growth in its revenue run rate – an extrapolation of full-year sales.

No bullish augury

The post-market performance of US IPOs in the 2026 vintage – down 4 percent while the Nasdaq 100 has risen 20 percent – is no bullish augury for Anthropic’s public offering. Nor is a 24 percent fall in Goldman Sachs shares since their July peak, given that the bank is lead underwriter for the biggest IPO jackpot in Wall Street history.

The political milieu in the US has also changed as a diesel shock and 7.5 percent 30-year fixed-rate mortgages trigger a Democratic Party blue wave in the November midterm election.

A public backlash against data centre construction is likely to dominate the legislative agenda of the next Congress, as will guardrails to regulate AI innovation after a dismal series of hacking incidents by rogue AI agents worldwide.

Moreover, a Federal Trade Commission investigation into Anthropic and OpenAI consumer disclosure policies may be amplified if the midterms weaken the Trump White House.

Even Anthropic’s CEO, Dario Amodei, has publicly expressed fears about the risk of a catastrophic event engineered by rogue AI agents and called for an immediate slowdown in AI innovation, while Bill Gates has warned that AI could cause a billion human deaths. This is not the narrative soundtrack Anthropic needs on the eve of an IPO at a time when risk and liquidity premia are rising alarmingly in financial markets.

The devil is in the aftermarket performance risk in the most-hyped AI IPO ever

GCC sovereign wealth funds and institutional investors were among the most prominent financiers of Anthropic in its early capital-raising rounds. Due to the scale of reconstruction and missile defence needs caused by Iran war damage, these funds need to raise liquidity once their investor lockups expire next year to repatriate petrodollars back to the Gulf.

This is a clear signal that Anthropic will find it difficult to sustain a $2 trillion market cap even if the IPO succeeds in mid-November. The devil is in the aftermarket performance risk in the most-hyped AI IPO ever.

Anthropic’s valuation metric at 435 times 2025 revenues makes me highly sceptical about the sustainability of its initial $2 trillion IPO target, no matter how stellar its growth has been in the past two years.

Buyer beware

Even the merest hint of an AI revenue slowdown, let alone “higher for longer” US Treasury bond yields, is likely to have a devastating impact on the present value of future discounted cash flows. This means a certain steep fall in the share price, as happened to SpaceX a month after its IPO, when the stock lost a trillion dollars in market cap on the Nasdaq.

Another major IPO risk is Anthropic’s staggering $518 billion in compute commitments to companies including Amazon, Google, Microsoft and SpaceX. This level of spending is difficult to justify, especially because 80 percent of the contracts are non-cancellable and may become catastrophic if AI product sales slow.

Customer concentration is another risk that may prove detrimental. Google, Amazon and Microsoft are Anthropic’s largest clients as well as its biggest investors. Any spending cuts by the Big Three hyperscalers will amount to a public execution for Anthropic shares.

And then there is governance. Anthropic is ostensibly a public benefit corporation but control is firmly in the hands of the founders, who retain 51 percent voting power via a special class of shares not available to public shareholders. Despite Amodei’s moralistic public pronouncement on ethics, the governance DNA of his AI firm’s IPO is no different from most Silicon Valley Big Tech hyperscalers.

The IPO’s postponement is itself a risk, as the CEO has recently prioritised the AI safety debate and White House photo opportunities with President Donald Trump over travelling with his investment bankers to market the offering.

My bottom line for this $2 trillion mega IPO? Caveat emptor.

Matein Khalid is an investor in global financial markets and board adviser to leading family offices in the UAE and Saudi Arabia

Register now: It’s easy and free

AGBI registered members can access even more of our unique analysis and perspective on business and economics in the Middle East.