Business WWC APAC Desk

Anthropic AI Valuation Faces a Profit Test—Can Explosive Revenue Growth Justify Investor Expectations?

Anthropic AI Valuation Faces a Profit Test—Can Explosive Revenue Growth Justify Investor Expectations?

SYDNEY / SAN FRANCISCO — AustralianSuper’s chief investment officer Shaun Manuell has raised a fundamental question about the artificial-intelligence investment boom: whether enormous spending on computing and technology will produce sustainable returns. As of Nov. 17, 2026, Anthropic’s own preliminary public-offering prospectus offers an unusually detailed test of that question, showing rapid revenue growth alongside substantial infrastructure expenses.

The Anthropic AI valuation debate is not simply about whether customers want AI. Its financial disclosures show demand, a large funding base and continuing costs. Investors must decide whether these justify market prices.

Anthropic AI valuation meets published financial figures

Anthropic’s preliminary prospectus, dated Nov. 10, sets out a proposed offering of 825 million Class A shares, including shares offered by existing holders, at an estimated $70 to $78 per share. The indicated range is preliminary and does not establish a final offering price or guarantee that the transaction will close.

The same document reports $17.0 billion of revenue for the quarter ended Sept. 30, 2026. It also reports a $567 million operating loss for that quarter. These figures can both be true: customer demand may expand extremely quickly while the cost of supplying AI services continues to absorb substantial resources.

The preliminary prospectus documents sales, expenses and investment obligations, subject to its stated qualifications.

AustralianSuper puts investment returns under scrutiny

AustralianSuper formally appointed Manuell as its chief investment officer effective July 1, 2026. In a published discussion with fund members, he described AI as a transformative technology but asked what returns would ultimately result from the scale of spending across the industry.

He also identified uncertainty over how spending on data centres and technology would become sustained earnings growth. This was a general investment assessment, not a documented allegation that Anthropic had manipulated its finances.

Transformative technology does not guarantee attractive investment returns.

That concern connects with earlier WWC coverage of China AI investment concentration, where the central issue is the allocation of capital rather than whether artificial intelligence can work.

Revenue growth does not settle the profitability question

Anthropic’s February 2026 financing announcement reported a $380 billion post-money valuation and a $14 billion annualised revenue run rate. Its May announcement reported a subsequent $65 billion funding round at a $965 billion post-money valuation, with run-rate revenue exceeding $47 billion earlier that month.

These are company-announced financing and run-rate measures. A run rate extrapolates a recent level of business activity; it is not the same thing as audited annual revenue. Similarly, a private financing valuation reflects the terms of a transaction, not a guaranteed future public-market price.

The preliminary prospectus supplies another perspective: it records $17.0 billion of revenue in just the third quarter, but computing and infrastructure expenses reached $9.9 billion during that period. The company explains that much of its compute is secured well ahead of use, leaving substantial near-term costs that cannot quickly be reduced.

Analysis: The useful metric is not merely how many customers adopt an AI tool. It is how much economic value remains after the company pays for the computing, distribution, employees and other services necessary to deliver it.

Infrastructure commitments create long-term exposure

Anthropic has announced major agreements with technology suppliers. In April it described an expanded arrangement with Google and Broadcom for next-generation computing capacity beginning in 2027, and separately announced an agreement with Amazon involving up to five gigawatts of additional capacity.

Those arrangements are relevant to future service delivery, but planned capacity is not synonymous with immediate revenue. Customers must use the capacity, and the economics must justify its cost.

The prospectus describes material future commitments and warns that computing arrangements can entail fixed or advance obligations. Investors therefore need to examine the timing of expenses against the timing of revenue and cash generation.

This is also why interconnected financing across the AI industry deserves scrutiny. A commercial partner can simultaneously provide capital and sell essential services; such relationships are not inherently improper, but they can complicate assessments of independent demand.

What an AI bubble comparison can and cannot establish

Analysis: Comparisons with earlier technology booms should be treated as questions about prices and expectations, not forecasts of an inevitable collapse. A new technology can become essential while investors still overpay for some companies or infrastructure projects.

Anthropic’s customer and revenue disclosures offer tangible evidence of commercial demand. Its operating loss and computing commitments show why demand alone does not establish a sustainable return on capital.

Retirement funds have a particular reason to examine this difference. Members rely on investment performance over many years, while an exciting technology investment can remain financially difficult even when its products are useful.

WWC has separately examined lending risks in AI infrastructure. The question is whether income covers financial commitments.

What happens next

Anthropic’s preliminary prospectus says it has applied to list its shares on Nasdaq under the proposed symbol ANT. The application and indicated price range are not proof that the IPO has been completed.

The next meaningful evidence will include the final offering terms, any updated disclosures and subsequent results showing whether revenue growth continues to outpace operating and infrastructure costs. Investors should also examine whether computing efficiency gains persist as the company expands capacity.

For now, the records support a balanced conclusion: Anthropic has reported extraordinary commercial growth, and AustralianSuper’s publicly documented concern about the returns on industry-wide investment remains relevant. Neither fact proves that all AI valuations are justified or that a market-wide crash is inevitable.

FAQ

Who is AustralianSuper’s chief investment officer?

Shaun Manuell became AustralianSuper’s chief investment officer on July 1, 2026, following the retirement of Mark Delaney.

What is Anthropic’s latest disclosed quarterly revenue?

Anthropic’s preliminary prospectus reports $17.0 billion of revenue for the quarter ended Sept. 30, 2026.

Is Anthropic profitable?

The prospectus reports a $567 million operating loss for the third quarter of 2026. That result should not be confused with revenue growth or other adjusted metrics.

Has Anthropic completed its IPO?

The Nov. 10 prospectus is marked preliminary and subject to completion. The record reviewed does not establish a completed offering.

Does high AI spending prove an investment bubble?

No. Large investments may ultimately generate lasting returns, but investors must evaluate profitability, financing requirements and valuation rather than assume demand will automatically translate into profits.

Get our stories first on Google

More in Business

See all in Business