New YorkUS edition · Updated hourlyLoading the wire
The intelligence brief for hedge funds, allocators and active investors

HedgeWeekly

Wall Street · Washington · The World
New York edition
Latest
Headlines appear here as the desk files
The first edition is being filed.The news desk publishes every hour. Stories appear as soon as the editor approves them.

Most recent

    Today's coverage

    US 0Global 0
    Rolling 24 hours · line marks the 60% US target

    HedgeWeekly Video

    The 60-second briefing · Weekdays before the bell
    Coming soonThe morning in sixty secondsThree stories moving US markets, filed before the open.
    Coming soonHedge fund positioningWhat the prime books are telling us this week.
    Coming soonEarnings, explainedThe numbers that mattered and what changed.
    SponsoredPresent this seriesTitle sponsorship of HedgeWeekly Video is open.

    Markets

    Tech & AI

    Anthropic's draft IPO prospectus shows $16.3 billion in first-half revenue and more than $600 billion of compute commitments

    A draft filing reviewed by The Wall Street Journal shows sales up roughly 15-fold from a year earlier, alongside an $80.7 billion net loss driven mostly by accounting charges.

    HedgeWeekly Tech & AI Desk ·

    Anthropic, the developer of the Claude AI models, generated about $16.3 billion of revenue in the six months to June 30, up from about $1.1 billion a year earlier, according to a draft initial public offering prospectus reviewed by The Wall Street Journal and reported on October 20. Anthropic declined to comment on the report. The document is still a draft, so figures could change before any public filing.

    The draft shows a first-half net loss of $80.7 billion, against about $13 billion a year earlier. About $72 billion of that came from non-cash accounting charges linked to changes in the value of convertible notes and other financing instruments, leaving roughly $8.7 billion. Spending on computing capacity and infrastructure, the processing power used to train and run AI models, rose to nearly $9 billion from $2.3 billion a year earlier, more than the roughly $7.3 billion spent in all of 2025.

    Anthropic's infrastructure commitments exceed $600 billion, according to reports on the draft. They include about $450 billion of mostly non-cancellable cloud payments over seven to 10 years, among them roughly $111 billion with Google, $110 billion with Amazon Web Services and $31 billion with Microsoft. They also include up to $86 billion to rent Nvidia-based capacity from Nscale and Lambda, and more than $160 billion in lease obligations for Google chips through Broadcom. Deals still under negotiation could add up to $150 billion. Two unnamed customers accounted for about a fifth of first-half revenue.

    Why it matters: the draft gives investors a first detailed look at the economics of a leading AI lab, and at how much of the cloud and chip industry's future demand rests on a single buyer. Those commitments underpin AI spending forecasts at Alphabet, Amazon, Microsoft, Broadcom and Nvidia. Customer concentration and the size of non-cancellable obligations are likely to be the focus as the filing moves towards a public prospectus. Anthropic separately told investors that third-quarter revenue reached $16.8 billion, according to earlier reports, but that figure has not been reconciled with standard accounting results.

    Sources

    The Archive