Artificial intelligence startup Anthropic recorded an $80.7 billion net loss across the first six months of 2026 as it prepares for an initial public offering, according to leaked draft regulatory filings reported Tuesday by The Wall Street Journal.
The massive deficit for the maker of the Claude chatbot grew sharply from a $13 billion loss recorded during the identical timeframe a year prior. However, the bulk of the shortfall stemmed from non-cash items, with roughly $72 billion tied to accounting adjustments connected to changes in the valuation of convertible notes and related financial instruments, the Journal reported.
Despite the steep bottom-line loss, Anthropic demonstrated rapid top-line growth. Revenue hit $16.3 billion for the six-month period ending June 30, surging from $1.1 billion during the first half of 2025. That expansion has been powered substantially by developer demand for Claude Code, its automated programming tool, Al Jazeera reported. More recently, company leadership informed potential backers that third-quarter sales reached $16.8 billion, a 45 percent increase from the preceding quarter, TechCrunch reported.
Staggering infrastructure demands
The disclosures illustrate the extraordinary capital requirements confronting advanced software developers. Anthropic indicated in its paperwork that it could pour more than $750 billion into cloud contracts in the coming years, including agreements that remain under negotiation, according to the Journal and TechCrunch. Meanwhile, Crypto Briefing reported the firm has earmarked upward of $600 billion directly toward computing resources.
Cash outflows for computing hardware and facilities almost quadrupled during the first six months of the year, nearing $9 billion, according to the Journal. Securing sufficient capacity has simultaneously grown more fraught. Leaked S-1 documentation revealed the startup is encountering difficulties obtaining enough cloud infrastructure from partners such as Amazon Web Services, Alphabet’s Google and Microsoft, who increasingly view Anthropic as a rival, TechCrunch reported.
To bypass these bottlenecks, Anthropic has turned to leasing proprietary data facilities and enlisting independent operators. Yet the filing noted concerns regarding those outside vendors, warning that their comparatively short operational track records could introduce reliability hazards compared to established hyperscale tech giants, according to TechCrunch.
Valuation pushback and catastrophic risks
The mounting expenditures have unsettled some financial backers. Investors attending an internal presentation in San Francisco last week began re-evaluating the business at approximately $1.5 trillion, Reuters reported, down from earlier expectations approaching $2 trillion. Prediction markets similarly signaled lowered odds that the firm will sustain its initial valuation benchmarks, according to Crypto Briefing.
Revenue concentration also presents near-term vulnerability. Only two corporate customers generated roughly one-fifth of Anthropic’s sales over the first half of 2026, the Journal reported.
Alongside financial constraints, the draft paperwork outlined severe potential operational risks. Anthropic warned regulators that damages resulting from rogue or harmful behavior by its models could ultimately outstrip its insurance coverage, TechCrunch reported. In additional regulatory risk disclosures, the company acknowledged unresolved challenges in safeguarding models against misuse in sophisticated extortion or digital attacks, while separately noting findings that software systems may soon acquire the ability to duplicate themselves across consumer electronic hardware, Al Jazeera reported.
Anthropic, which Reuters reported logged a $42 billion net loss throughout 2025, has not publicly commented on the disclosures. Market observers are watching whether upcoming discussions with strategic cloud partners or formal revisions in regulatory filings will stabilize investor sentiment before its planned public debut.
