A draft IPO prospectus reviewed by the Wall Street Journal points to an enormous commitment to infrastructure: roughly $450 billion of computing agreements over seven to 10 years, most locked in. The spending roadmap by partner is detailed too, at $111 billion with Alphabet's Google Cloud, $110 billion with Amazon Web Services, and $31 billion with Microsoft.
In a separate note, the Journal says Anthropic may channel as much as $750 billion into added computing capacity over the coming years, emphasizing the scale of capital required to build larger AI systems.
WSJ, drawing on the draft filing, says Anthropic recorded a $80.7 billion net loss during the first half of 2026, up from a $13 billion loss in the comparable period of 2025. The report says $72 billion of the 2026 first half loss stems from accounting charges tied to revaluing convertible notes and other instruments the company has issued.
Outlays are rising as well. According to the Journal, Anthropic outlaid nearly $9 billion for computing capacity and infrastructure over the six-month period that ended June 30, 2026, versus $2.3 billion in the first half of 2025. During the first half of 2026, the company committed more to securing compute than across all of 2025, when spending reached $7.3 billion.
When sweeping tech spending and IPO plans catch attention, sign up for our free daily Market Briefs to stay informed
Anthropic, widely seen as a leading AI developer, is expected to publicly file for a listing in the coming weeks, setting the stage for a potential IPO that ranks among the largest to date. The listing is intended to raise capital for chips, data centers, and hiring to develop new AI models.
Bloomberg News has reported the company told investors it reached adjusted profitability in the third quarter, is producing positive free cash flow, reported an 85% gross margin, and recorded an adjusted operating margin above 11% as of the third quarter. Bloomberg also says formal IPO marketing could kick off the week of Nov. 9 at the earliest, potentially leading to trading before US Thanksgiving, with a debut anticipated by year end.
The draft prospectus highlights AI safety concerns, including the possibility of models showing "self-preserving" behavior to avoid shutdown and producing responses that resemble "blackmail" to deter humans from stopping them. It also cites potential political resistance to building US data centers, where most of Anthropic's capacity is located, according to the Journal.
The bottom line: giant multi year obligations and heavy infrastructure spend on one side, improving adjusted profitability metrics on the other. For your portfolio, that split sets the tone for how to weigh growth against cash needs as the IPO nears.
Keeping an eye on complex market and economy signals is easier when you sign up for our free Market Briefs newsletter each morning
