Anthropic IPO Draws Skepticism From Australian CIOs

NextFin News - Australia's most influential superannuation investors are balking at the valuation of Anthropic's blockbuster initial public offering, a rare moment of pushback from a pool of capital that has spent the past decade chasing growth at almost any price. The skepticism from Australian chief investment officers lands as the Claude creator files to list on Nasdaq at a valuation that has been discussed as high as $2 trillion, setting up what independent research firm New Constructs has already called the "most ridiculous IPO of 2026."

The tension is stark. Anthropic's annualized revenue run rate hit $65 billion at the end of July, a sevenfold increase from the end of 2025, according to figures confirmed by people familiar with the matter. At a $2 trillion listing, that is roughly 31 times revenue. Even at the lower end of what investors now privately expect — between $1.4 trillion and $1.75 trillion — the company would still price at more than 20 times a run rate that may be slowing as the AI funding well runs dry.

For Australian investors, the question is no longer abstract. Anthropic has filed a prospectus with the Australian Securities and Investments Commission, using the same brokerage arrangement that gave retail investors access to Elon Musk's SpaceX float in June. The offering is coming to a market that is still recovering from the collapse of Firmus, a data-center deal that failed at a $44 billion valuation earlier this year. Australian superannuation CIOs, who control the retirement savings of millions, are being asked to decide whether the AI trade has one more leg left — or whether they are being invited to buy at the top.

The Numbers Behind the Doubt

The skepticism is rooted in arithmetic, not mood. Anthropic reported revenue of $4.6 billion for all of 2025 and a net loss of $42 billion. Its first-half 2026 revenue reached $16.3 billion, up from $1.1 billion in the same period a year earlier, but the company's headline net loss for the period swelled to $80.7 billion. Of that, approximately $72 billion is an accounting artifact: the company's own convertible notes appreciated in value as its private valuation rose, creating a paper loss that does not reflect cash operations. The operating loss was $693 million, and the company generated $3.9 billion of positive operating cash flow in the same period.

That distinction matters to the bulls, and it does narrow the gap — but it does not close it. To justify a $2 trillion valuation at a 10 percent cost of equity, a 25 percent free-cash-flow margin and a 25 times terminal multiple, Dr. Chan Ahn, founder of Tessera PE and a former Goldman Sachs and JPMorgan executive, estimates Anthropic would need roughly $725 billion in annual revenue by 2036. At a 13 percent discount rate, the required revenue climbs to about $950 billion. For scale, Nvidia — the world's most valuable technology company — generated just over $190 billion in net income over its trailing four quarters. Anthropic would need to record roughly double that in annual profit to support the targeted price.

New Constructs, an independent research provider, values Anthropic's public equity at about $150 billion, implying roughly 93 percent downside from the targeted IPO valuation. The firm described the offering as an "unprecedented test of investor gullibility," a phrase that has gained traction among institutional investors who watched the 2021 software boom turn to dust.

Michael Burry, the investor known for his bet against subprime mortgages, added his own comparison: Anthropic's valuation equals the combined market value of 78 S&P 500 companies, including household-name brands and established industrial firms. The point is not that Anthropic cannot grow. It is that the growth required to justify the price has no precedent in public-market history.

There is one more risk buried in the filing that the headline revenue multiple does not capture: customer concentration. Roughly one-fifth of Anthropic's revenue came from just two customers, with the remainder coming from a small handful of clients. A company whose demand is concentrated in a few counterparties is more exposed to a single procurement decision than the run-rate multiple suggests.

Why Australian CIOs Are the Canary

Australian superannuation funds are not natural skeptics. They are among the world's most aggressive allocators to unlisted growth assets, with a structural mandate to chase returns over decades-long horizons. That is precisely why their hesitation is significant. When the investors who built their reputations on buying early and holding through cycles start questioning an entry price, the marginal buyer is running out.

Roger Montgomery, a prominent Australian fund manager, put the timing bluntly.

Anthropic is trying to pull off a historic IPO right as the AI funding well is completely running dry.

I think they're trying to get ahead of the bubble possibly bursting.

The reference point for Australian investors is fresh and painful. Earlier this year, local superannuation CIOs were widely credited with helping to shield the Australian sharemarket from the Firmus float, which collapsed after investors baulked at its valuation and the viability of its data-center model. The irony has not been lost on local commentators: the same CIOs celebrated saving the domestic market from an overvalued infrastructure deal, yet few have been willing to challenge the AI positions accumulating in their global equity books.

John Pearce, chief investment officer of UniSuper, one of Australia's largest superannuation funds, has struck a more measured tone but has not dismissed the risk. In an interview, Pearce said:

We are seeing a lot of investment in AI. I think valuations are looking a bit stretched and there are signs of frothiness but there is no certainty that we're in a bubble, and you actually don't know whether you're in a bubble until the bubble bursts.

Dan Farmer, chief investment officer of the $86 billion MLC Super, warned in October that the "AI juggernaut" could face a "vicious" correction. Cameron Gleeson, head of investment strategy at Betashares, acknowledged the other side of the trade:

Anthropic's growth puts it in rarefied air.

The Bull Case: A Watershed for the AI Revolution

The bulls are not arguing about multiples. They are arguing that Anthropic is the central infrastructure play of a technological shift that will redefine the global economy, and that conventional valuation frameworks do not apply to a company sitting at the chokepoint of the AI buildout.

Dan Ives, a technology analyst at Yorkville Ives, called the listing

a watershed event, not just for tech, but the AI revolution in terms of the role Anthropic plays.

Thomas Hayes, chairman at Great Hill Capital, framed the offering in almost ideological terms.

The offering is for the faithful, that sooner or later, they're going to get some return on that invested capital.

There is evidence for each pillar. Anthropic's second-quarter revenue of $11.5 billion was a 14-fold jump from a year earlier. The company has signed approximately $518 billion in multi-vendor compute and infrastructure commitments, locking in the capacity it needs to meet surging demand. And its Claude models have become the default choice for a growing share of enterprise and developer users, displacing OpenAI in several adoption metrics. On an adjusted basis that strips out the convertible-note charge, the company reported adjusted net income of $2.7 billion for the first half of 2026 — the number the bulls point to when they say the losses are mostly accounting.

But the bull case has a timing problem. Matt Kennedy, a senior strategist at Renaissance Capital, which provides IPO-focused research, noted that

Anthropic is entering an IPO market that is far more challenging than the summer. Investors were not questioning valuations the same way a few months ago.

Cyclical or Structural: The Call That Decides the Trade

Underneath the debate over multiples is a deeper question: is the skepticism about Anthropic a cyclical reaction to a tighter funding environment, or a structural judgment that the AI buildout itself is overbuilt?

The evidence points to both, operating on different time horizons. In the short term, the pressure is cyclical. Interest rates are higher than they were in the summer, debt is harder to raise, and the marginal buyer of IPO paper is more discriminating. Cyclical pressure is mean-reverting: if rates fall and liquidity returns, valuations can reflate without any change in fundamentals. A rate cut, or a fresh wave of sovereign-wealth demand for AI exposure, would be enough to carry the offering.

But the structural question is harder to dismiss. The AI buildout is capital-intensive in a way that previous technology cycles were not. Anthropic's $518 billion in compute commitments is not a one-time cost; it is a recurring obligation that must be serviced regardless of whether demand keeps pace. If enterprise adoption plateaus before revenue catches up with capacity, the company faces a margin squeeze that no multiple expansion can offset. That is a structural risk, and it does not mean-revert on its own.

The distinction matters for Australian investors in particular. Superannuation funds hold assets for decades, which means they can afford to wait out a cyclical trough. They cannot afford to be wrong about a structural shift. If the AI buildout is a durable regime change, buying at 30 times revenue today will look cheap in five years. If it is a capital cycle that has overshot, the same purchase will anchor a decade of underperformance.

The Second-Order Risk: Who Is Left Holding the Compute

The first-order consequence of a failed or disappointing Anthropic IPO is obvious: the stock falls, and the investors who bought at the offer price lose money. The second-order consequence is more dangerous, and it is the one the market has not fully priced.

Anthropic's $518 billion in compute commitments is not just Anthropic's problem. It is a claim on the balance sheets of the cloud providers, chipmakers, and lenders who underwrote the buildout. If Anthropic's revenue growth slows, the company will have an incentive to renegotiate or walk away from those commitments — exactly the dynamic that has already pushed more than 200 neocloud providers into insolvency. A disorderly Anthropic debut would force the market to reprice the entire chain of AI infrastructure financing, from Nvidia's contracted backlog to the debt facilities that funded data-center construction.

This is why the Australian CIOs' skepticism is more than local color. Australia's superannuation system is one of the world's largest pools of institutional capital, and its CIOs have been among the most willing to allocate to unlisted technology and infrastructure. If they begin to treat AI exposure as a risk to be managed rather than a theme to be chased, the flow of capital into the AI buildout could slow at precisely the moment the companies in the chain need it most.

The counter-thesis is straightforward and deserves a fair hearing: the skeptics have been early before. Every transformative technology — the internet, mobile, cloud computing — produced a valuation bubble on the way to a much larger destination. The investors who waited for "reasonable" multiples in the late 1990s missed the bulk of the returns in the following decade. Anthropic's revenue is growing faster than any company in history at a comparable scale, and the total addressable market for frontier AI may be an order of magnitude larger than the most optimistic modelers assume. If that is true, today's skeptics will look like the investors who called Amazon expensive in 2005.

The falsifying signal is specific. If Anthropic's annualized revenue run rate fails to reach $100 billion by the end of 2026 — the level many investors already expect — or if the company's operating margin does not improve materially from its current negative position within four quarters of listing, the structural-bull thesis is broken. At that point, the valuation would rest entirely on multiple expansion, and the IPO would be a capital-cycle top rather than a generational entry point.

What to Watch

The offering is expected to price in the fourth quarter of 2026, with November the most likely window after the U.S. midterm elections. Three signals will determine whether the skeptics or the faithful are right.

First, the final pricing. A listing at the lower end of the $1.4 trillion to $1.75 trillion range would signal that underwriters have absorbed the skepticism and are leaving money on the table for investors. A price near $2 trillion would signal that demand remains strong enough to ignore the arithmetic.

Second, the first-day and first-month trading performance. The SpaceX float rose from $135 to $161, a 19 percent gain, on its debut. Anthropic will be judged against that benchmark. A weak debut would confirm that the marginal buyer has indeed become more discriminating.

Third, the response of the Australian superannuation funds. If the large CIOs limit their exposure or sit out the retail tranche entirely, it would be a quiet but powerful signal that the institutions best positioned to hold through volatility have decided the price is wrong.

In the end, the Anthropic IPO is a test of a simple proposition: whether the market will pay today for a future that has never been demonstrated at scale. The Australian CIOs, burned by Firmus and wary of the drying funding environment, are signaling that they need proof before they commit. The faithful, led by Ives and Hayes, are asking investors to look past the numbers and bet on the revolution.

History suggests both sides will find evidence for their view. The investors who get this right will be the ones who recognize that a transformative technology and a fair price for its stock are two different questions — and that answering the first correctly does not guarantee a profit on the second.

Explore more exclusive insights at nextfin.ai.