MACRO FIRESIDE | OCTOBER 26, 2026
Sequel to “My first impressions,” published earlier today (10/26/26)
Diving a little deeper into Anthropic’s S-1, I find the infrastructure commitments far more intriguing than the $80.7 billion net-loss headline. As I argued earlier today, that headline tells us surprisingly little about the operating business.
Claude may be sold as software, but delivering it requires enormously expensive physical infrastructure. Chips, data centers and power cannot be procured off the shelf. Anthropic says it generally secures compute 12 to 18 months in advance.
Three numbers stood out to me in particular. As of June 30, Anthropic reported $265.6 billion in future fixed or minimum purchase commitments, and $321.4 billion in estimated payments for leases signed but not yet commenced. New and amended agreements after June carried approximately $245 billion in stated minimum obligations. The arrangements differ in duration and terms, and amended contracts add another layer of complexity. Indeed, the filing leaves the aggregate commitment figure blank. Granted, these obligations are not all debt, and a simple addition of the figures could be misleading. But together they do represent enormous potential claims on future cash flows. After all, the obligations must ultimately be met without any assurance about customer demand and pricing years down the road.
That being said, Anthropic is already getting far more revenue out of each dollar spent on compute. Compute and infrastructure costs, as a percentage of revenue, fell from 201% in the first half of 2025 to 55% in the comparable 2026 period, as noted earlier. The company generated $3.9 billion in operating cash flow during the half, although some of that reflected favorable working-capital movements rather than cash earnings alone. Anthropic also held $131.7 billion in cash and short-term investments as of June 30. Financing cash inflows alone were $110.8 billion during the half. That financing flywheel may have to keep turning for the foreseeable future, it appears.
Perhaps what concerns me more than the sheer size of these commitments is the mismatch between relatively fixed future costs and largely variable revenues. For instance, four-fifths of first-half revenue was consumption-based. Moreover, many of Anthropic’s largest customers are not bound by long-term contracts, while its top five customers accounted for roughly a quarter of revenue over the first nine months. A 900% net revenue retention is extraordinary. That is to say, among the qualifying customers included in the measure, quarterly revenue was roughly nine times what it had been a year earlier. But, to me, it does not seem like a sustainable rate of expansion.
Cheaper compute may expand the market for the firm, but is it necessarily better for Anthropic shareholders? I wonder. Customers will demand some of the savings, and competition may force the company to concede more than it would like. Rising volumes might more than compensate, but we have yet to observe the pricing and margins of this industry through a normal competitive cycle.
Investors are discussing a $1.5 trillion valuation. That works out to roughly 32 times annualized June-quarter revenue, not a forecast and not the price in the prospectus. I would be careful about putting a software multiple on a business with such large, multi-year capacity commitments.
At the end of it all, what I want to know is simple: how much of the next dollar of revenue will be left for shareholders once the machines have been paid for?
Disclaimer
The Macro Fireside is published for informational and educational purposes only. Nothing in this publication constitutes investment advice, a solicitation, or an offer to buy or sell any security, financial instrument, or investment product of any kind. The views expressed are solely those of the author and do not represent the views of any employer, affiliated entity, or counterparty.
All analysis reflects the author’s independent judgment as of the date of publication. Market conditions, data, and regulatory environments change rapidly; no representation is made that any information herein remains current or accurate after the publication date. Past performance of any asset class, strategy, or instrument referenced herein is not indicative of future results.
Readers should conduct their own independent research and due diligence, and consult a qualified financial adviser, attorney, or tax professional before making any investment decision. The author may hold positions in securities or instruments discussed in this publication. Such holdings are subject to change at any time without notice and without obligation to update this publication.
This publication is not directed at, and should not be relied upon by, any person in any jurisdiction where its distribution or use would be contrary to applicable law or regulation. By reading this publication, you acknowledge and agree that the author and The Macro Fireside bear no liability for any investment decisions made in reliance on its contents.