Anthropic’s $80 billion loss explained: Why its AI business still faces a big test

Anthropic’s reported loss of more than $80 billion in the first half of 2026 comes with a crucial distinction. Most of it stemmed from accounting charges linked to the value of financial instruments it issued to raise funds, leaving investors to assess a different question: whether its booming AI business can sustain the cost of securing computing power.

The artificial intelligence company recorded a net loss of $80.7 billion in the six months ended June 30, compared with $13 billion in the same period a year earlier, according to a draft initial public offering (IPO) prospectus cited by The Wall Street Journal. About $72 billion of the loss came from accounting charges reflecting an increase in the estimated value of bonds that could eventually convert into Anthropic shares.

These charges do not represent an equivalent amount of cash spent on running the business. However, they have a significant impact on its reported financial results as Anthropic prepares for a potential stock market listing.

Why Anthropic’s loss needs context

Convertible bonds are financial instruments that can turn into company shares under agreed conditions. Changes in their estimated value can create accounting losses for the company issuing them, depending on the terms and applicable accounting rules.

In Anthropic’s case, the $72 billion charge accounted for nearly 90 per cent of its reported first-half loss. The figure should therefore not be interpreted as $80.7 billion in cash burned during the period.

More from Business

But the accounting distinction does not eliminate concerns about the company’s underlying finances. Reuters reported in September that Anthropic recorded an operating loss of about $8.1 billion in 2025, excluding the large accounting charges associated with its financing instruments.

An operating loss occurs when a company’s recognised operating expenses exceed its operating revenue. It provides a clearer indication of the costs of running the business, although it does not by itself establish how much cash the company has spent.

Revenue surges as infrastructure costs climb

Anthropic’s revenue rose to $16.3 billion in the first half of 2026 from $1.1 billion a year earlier, according to the WSJ report. Its spending on AI infrastructure nearly quadrupled to almost $9 billion during the period, compared with $2.3 billion in the first half of 2025, Reuters reported.

The company has benefited from demand for Claude, its AI assistant, and Claude Code, a coding tool used by developers to write software and build applications.

Meeting that demand requires substantial computing capacity. AI companies need specialised chips, data centres and cloud services to train models and run them for customers. As usage grows, they must secure additional capacity, creating significant infrastructure costs and long-term financial commitments.

Anthropic has indicated that it could spend as much as $750 billion over the coming years to secure computing capacity, including potential deals under discussion, the WSJ reported.

Separately, Reuters reported that the company’s draft prospectus detailed at least $518 billion in commitments over a decade for cloud services, computing and infrastructure. The figures reflect different scopes of spending and should not be added together.

What investors will watch ahead of the IPO

Anthropic’s growth remains strong, but the pace is becoming an important consideration for investors. The company told prospective IPO investors that third-quarter revenue rose 45 per cent from the previous quarter to $16.8 billion and that it was profitable on an adjusted basis, Reuters reported.

The increase followed a period of particularly rapid expansion. Revenue more than doubled between the first and second quarters, reaching $11.6 billion in the latter. Some investors were disappointed by the slower sequential growth rate, Reuters reported.

Adjusted profitability is not the same as net profit under standard accounting measures. Investors will need to assess the underlying costs of serving customers, future infrastructure commitments and the company’s ability to finance expansion.

The disclosures have also prompted fresh scrutiny of Anthropic’s potential valuation. Some current and prospective investors have discussed a valuation of around $1.5 trillion, below the roughly $2 trillion previously expected for its planned November share sale, Reuters reported.

The IPO could become an important test of investor appetite for large AI companies ahead of a potential public listing by rival OpenAI next year.