Anthropic logged net loss of over US$80 billion in first half of 2026: Report

Artificial intelligence developer Anthropic has recorded a staggering net loss exceeding US$80 billion during the first half of 2026, according to details revealed in a draft initial public offering prospectus. The figures illuminate the immense capital expenditure required to compete in the global artificial intelligence sector, even as top-line revenue surges on the back of products like the Claude coding assistant. The financial disclosures have triggered intense debate among prospective public market investors ahead of a planned share sale, arriving as the wider technology market scrutinizes the sustainability of heavy spending on specialized chips and data center infrastructure.

The vast majority of the first-half loss — specifically US$72 billion — stemmed from non-cash accounting adjustments and changes in the value of convertible notes and other financing instruments issued to raise capital, as reported by The Straits Times. This marked a dramatic escalation from the US$13 billion net loss recorded during the corresponding period a year prior, which itself followed a net loss of US$42 billion for the full year of 2025, according to reporting outlined by SiliconANGLE.

Media additions

Despite the headline losses, operational revenue scaled rapidly. Anthropic generated US$16.3 billion in revenue during the first six months of 2026, dwarfing the US$1.1 billion reported in the first half of 2025, according to investingLive. That performance follows a full-year revenue of approximately US$4.6 billion for 2025, which had already represented a substantial multiple over earnings from 2024.

The financial documents also detail monumental capital commitments for future computing capacity. According to The New York Post, Anthropic faces more than US$600 billion in future cloud computing and chip-leasing costs, with additional deals under negotiation that could inflate that figure by another US$150 billion.

Executive compensation figures within the draft prospectus revealed that co-founders Dario Amodei and Daniela Amodei drew modest base salaries of roughly US$655,000 each in 2025. Stock and option awards brought Dario’s total compensation package to about US$18 million, while Daniela’s package was valued at approximately US$16.5 million. Both executives have pledged large portions of their personal equity to philanthropic causes, mirroring broader corporate governance commitments such as a recent pledge to allocate five percent of shares to a public benefit fund.

Risk disclosures in the filing highlight several operational headwinds, including intense customer concentration where two clients account for roughly one-fifth of first-half revenue. Furthermore, Anthropic warned of a growing, organized political opposition to data center construction across various jurisdictions, noting that its high public profile could invite heightened scrutiny toward its infrastructure projects. Lior Prosor, a partner at Deep33 Ventures, told The New York Post that data centers currently entering the permitting process would not be operational before the second half of 2028, suggesting that immediate IPO prospects would remain insulated from near-term construction delays.

Market reactions to the leaked filings have been mixed. During a recent investor day in San Francisco, attendees discussed a potential valuation of around US$1.5 trillion for the upcoming share sale, pulling back from the US$2 trillion figure previously anticipated by the market, as noted by Central Oregon Daily. At the same time, company executives told investors that third-quarter revenue reached US$16.8 billion, representing sequential growth over the second quarter, and that the firm achieved profitability on an adjusted basis during that three-month window.

As preparations continue, financial analysts and market participants will monitor the formal progression of the initial public offering. The success or cooler reception of the share sale is widely expected to influence broader sentiment across the artificial intelligence sector and set valuation benchmarks ahead of potential competitor listings, such as rival OpenAI's potential public debut.