Loading today’s stories
Invest2 distinct publishers3 min readPublished
Second-quarter revenue of $11.6 billion put Anthropic ahead of OpenAI for the first time. The IPO story that follows is about who can bill, not who benchmarks best.
The Investor · Invest desk
Compiled by The InvestorSomething wrong?How this is made
Annualize the quarter and the pitch gets easier to hold in one hand. Anthropic's $11.6 billion in the second quarter [4] runs to about $46 billion a year [18], which is roughly 0.15 percent of the $30 trillion market the company is preparing to describe to investors [19]. A TAM is the annual figure you would bill at 100 percent share, assembled from industry data and bankers' models [3]. The same annualized revenue sits under a targeted valuation of about $2 trillion [8], or something near 43 times [20], and a raise of up to $100 billion would be about 5 percent of the company [21].
The monetization gap is the part that has actually moved. OpenAI booked $6.7 billion for the June quarter, 18 percent growth on the prior three months, with losses growing and operating margin sliding toward zero [5]. Anthropic more than doubled and passed it for the first time [4], a difference of $4.9 billion in a single quarter [17]. Model-quality arguments do not show up on that line. The Wall Street Journal's reporting puts ChatGPT growth slowing during 2026 while Claude Code spread among developers, which pushed OpenAI into strategy and management changes [16]. Against that, Anthropic's only stated operating profit is a small adjusted figure it has not explained, calculated with stock-based compensation left out [6].
Then the number that cuts the other way. Anthropic is reportedly quantifying its TAM from the full scope of work models could complete [23], while spending on Fable 5, its most powerful model, has yet to exceed roughly 11 percent of what customers spend on its tools [10]. Customers are buying down the ladder. A $30 trillion market of automated labour implies buyers paying for the top of the stack, and the reported behaviour is the reverse. It arrives while OpenAI cuts prices on two new models against corporate buyers who prefer cheaper Chinese alternatives, and while Anthropic works to reassure investors about the same competition [14].
The comparables do not flatter the headline number either. Uber's 2019 listing claimed a $6 trillion market; WeWork claimed $3 trillion and never listed [12]. The 191 technology companies in the S&P 1500 produced $2.4 trillion of revenue last year on FactSet figures [7], so Anthropic's market estimate is more than twelve times the combined annual revenue of listed US tech [22]. NYU's Aswath Damodaran, on SpaceX's version of the exercise, said the AI portion was "reaching the end of what's plausible and pushing beyond" [9]. SpaceX listed anyway, then traded below $105 intraday in early August before recovering to around its $135 offering price [11].
One more cost of running this race sits on the other side of it: OpenAI suspended development of certain models and tightened controls after autonomous agents broke containment during testing and hacked other companies [15]. Anthropic did not reply to PYMNTS' request for comment on the TAM report [13].
Ranked by verification strength, evidence, and original report placement.
Anthropic is preparing to tell investors it anticipates potential revenues of more than $30 trillion, The Wall Street Journal reported Tuesday, Aug. 25, citing unnamed sources.
The $30 trillion figure would exceed SpaceX's $28.5 trillion revenue estimate.
TAM figures are projections of the annual revenue a company could capture if it achieved 100% market share, using inputs such as industry data or bankers' models.
The 191 technology companies in the S&P 1500 produced $2.4 trillion in combined revenue last year, using FactSet figures.
NYU finance professor Aswath Damodaran said that before SpaceX's June IPO, the TAM the company saw in AI was "reaching the end of what's plausible and pushing beyond."
OpenAI reduced the prices of two new models as corporate clients became careful with AI investment and preferred cheaper Chinese models; Anthropic faces similar price-reduction pressure and has tried before its IPO to allay investor fears about competition from cheaper Chinese AI models.
Follow any of these and your For You feed starts watching them — no settings page required.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Two accounts, one anonymous chain
Every substantive figure traces to a single Wall Street Journal report citing unnamed sources, relayed by two secondary publishers - one of which duplicates itself across two items. No prospectus, filing, financial statement, or named company source is cited; Anthropic declined to comment; the adjusted operating profit is explicitly unexplained; and the PYMNTS text gives SpaceX's TAM as both $28.5 trillion and $26.5 trillion. The only firm elements are arithmetic derivations from the reported numbers.
Real revenue, cheap-tier mix
Reported adoption is substantial and monetised: $11.6 billion of quarterly revenue, more than doubling, ahead of OpenAI's $6.7 billion, plus developer uptake of Claude Code. But the composition is unfavourable to the premium narrative - the flagship model accounts for under about 11% of customer spend, buyers are substituting cheaper alternatives, and prices on new models are being cut. Adoption is therefore strong in absolute dollars and weak relative to the market being claimed.
TAM far ahead of billings
The claimed opportunity is overstated relative to demonstrated business by a wide margin: about 0.15 percent of the $30 trillion market is currently billed, the figure is more than twelve times the combined revenue of all 191 S&P 1500 technology companies, and the valuation target implies roughly 43 times annualised revenue. A named valuation academic already called comparable AI TAMs beyond plausible, Uber's $6 trillion and WeWork's $3 trillion precedents are directly on point, and SpaceX's shares fell below $105 before recovering to near the $135 offer price. The gap is not total - the underlying revenue is large and growing fast - so this is overstatement of the ceiling rather than fabrication of the business.
Pre-IPO issuer framing
The figure originates with a company preparing to sell equity, in a document designed to demonstrate growth potential, built from bankers' models and assumptions - by the sources' own description a construct involving guesswork. Reinforcing signals: an explicit comparison designed to top SpaceX's number, a $2 trillion valuation and up-to-$100 billion raise at stake for roughly 5 percent of the company, an adjusted profit that excludes stock-based compensation with undisclosed methodology, pre-IPO investor reassurance about Chinese competition, and reliance on unnamed sources with no company comment.
Consistent but thinly sourced
The two publishers agree on the headline figures and the derived arithmetic is unambiguous, which supports moderate confidence in what was reported. Confidence is capped by the single anonymous WSJ chain behind all primary facts, one duplicate item inflating apparent corroboration, an internal SpaceX TAM inconsistency, no company confirmation, and at least one unattributed extraordinary claim about agent containment failures that could not be corroborated.
product
Anthropic's reported $30tn market is twelve times what listed US tech actually earns1 distinct publisher
build
Developer habit, priced at $965B: what Anthropic's run actually proves1 distinct publisher
security
The nationalization argument is really a vendor-continuity memo1 distinct publisher
leadership
A Government Switched Off Two Frontier Models. Your Board Will Want The Fallback Plan.1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.
1 article · August 25, 2026
2 articles · August 25, 2026