Loading today’s stories
Leadership1 distinct publisher3 min readUpdated
Exponential View's bottom-up model puts trailing generative AI sales at $110 billion, with the latest month annualising above $175 billion. The exclusions decide whether your budget is comparable.
The Board Room · Leadership desk
Compiled by The Board RoomSomething wrong?How this is made
The deduplication rule is what makes the figure usable, and also what keeps it small. The model counts the dollar an end customer pays and stops there: a dollar spent with Anthropic on Claude, of which fifty cents goes to Amazon to serve it, is reported as one dollar rather than a dollar fifty [4]. Both legs are tracked internally, but only the end-customer dollar reaches the headline [4]. Anyone who has sized this market by adding up vendor revenues has been counting the same money more than once.
The arithmetic behind the two headline numbers is worth doing yourself. A $175 billion run rate implies a most recent month of roughly $14.6 billion [1]. Set that against the trailing twelve months and the run rate is about 1.59 times the full prior year of sales, meaning the latest month annualised sits roughly 59 percent above the trailing-year average month [2]. That single ratio is the growth claim. It is also the number to handle carefully, because a run rate is one month with a multiplier attached.
What the count leaves out matters more than what it includes if you are using it as a benchmark. Professional services and systems integration are excluded, on the reasoning that when a Fortune 500 company commits to AI, only part of that spend reaches an AI company [5]. Internal AI uplift, such as better recommendation systems lifting ad revenue at Meta or Google, is excluded, as are internal efficiency savings [6]. China is modelled but not in this version [7]. So a finance line that bundles integrators and change management is not measuring the same thing, and any share-of-market calculation against this denominator is a non-China calculation.
The report sets out to answer how far these revenues go towards covering the investment expense [14], and the supply side of that equation is the part already visible: chips, memory, power transformers and cooling sit largely in public companies whose disclosures and forward order books show what is being spent [13]. The coverage ratio is the division between the two sides. Now that the denominator exists, the absence of that quotient is the most conspicuous gap in the picture.
Confidence should be set by how the number was built. Much of the revenue sits in private companies including OpenAI, Anthropic, Cursor and ElevenLabs, none of which is obliged to disclose anything [9], and the rest flows to hyperscalers that are public but do not consistently break out AI segment revenue [10]. The model works from public statements by hyperscalers, neoclouds, suppliers and customers, plus well-reported leaks and self-reports carrying confidence scores [11], assembled into item-by-item P&L, balance sheet and cash flow reconstructions that the authors say are auditable back to the contributing data point and its weighting [12]. That is a stronger method than a survey and still not a filing.
The growth comparison, roughly three times faster than the mobile or internet waves [8], and the reported intent of senior executives across industrials, insurance, finance and pharma to invest more [15], both come from the same publisher. The first is model output. The second is Azeem Azhar's conversations, which is directional evidence, not a booked order.
Follow any of these and your For You feed starts watching them — no settings page required.
Ranked by verification strength, evidence, and original report placement.
Exponential View says the work took several months and is, as far as it knows, the first bottom-up, deduplicated measure of consumer and enterprise AI spending across the full stack, released in its first The State of the AI Economy report.
The model draws on public statements from hyperscalers and neoclouds, their suppliers and their customers, using only high-confidence detailed facts, plus well-reported leaks and self-reports to which a confidence score is assigned.
The result is an item-by-item financial model for the largest contributing companies and business units, each a deconstructed P&L, balance sheet and cash flow, triangulated against external sources and internal consistency checks, which the authors say makes the numbers auditable back to the data point and its confidence weighting.
Many companies have moved beyond occasional pilots but are still early in scaling and deepening, and in Azeem's conversations with senior executives across industries in Europe and the US, including industrials, insurance, finance and pharma, the consistent message is that they intend to invest more.
The generative AI economy generated $110 billion in sales over the past 12 months, on a deduplicated basis.
Annualising the most recent month's revenues indicates an AI revenue run rate exceeding $175 billion.
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.
Detailed methodology, single interested source, no reproducible data
The cluster rests on one article by the party that produced the research. The counting rules, input classes and exclusions are unusually well documented, which lifts evidence above bare assertion, but the model, per-company estimates and confidence weights are not published, no ranges accompany the aggregates, inputs include leaks and self-reports, and no independent publisher corroborates the numbers.
Large modelled spend, but estimated rather than disclosed
The story reports substantial real-money AI purchasing — $110 billion trailing and an accelerating latest month — which is meaningful adoption evidence, and it is corroborated qualitatively by executives saying they are past pilots. But every number is a third-party estimate of undisclosed revenue rather than vendor-reported figures, and the count omits professional services, internal uplift and China, so the measured level of actual buyer behaviour is directional rather than firm.
Mildly overstated by framing, tempered by disclosed limits
The first-of-its-kind framing and the 3x-faster-than-mobile comparison run ahead of what the supplied text substantiates, and precise single-point aggregates carry more apparent precision than confidence-scored leaks and self-reports can support. The gap stays modest because the publisher volunteers its exclusions, its China omission and the opacity of the underlying revenue in the same piece.
Publisher is launching the paid research product it is citing
The only source is Exponential View promoting the launch of its first The State of the AI Economy report and its proprietary model, so the outlet reporting the numbers directly benefits from their being seen as novel, rigorous and important. The methodology transparency is real but sits inside a launch post, and the piece does not acknowledge this commercial interest.
Low: one self-interested publisher, unreproducible aggregates
Claims about what the report says and how it was built are firmly grounded in the text, so descriptive confidence is high. Confidence in the magnitudes themselves is low: one publisher, who authored and profits from the research, supplies unreproducible point estimates without ranges, and no second source in the cluster confirms or disputes them.
build
SemiAnalysis to software teams: your token cost starts at the fab, not the price list1 distinct publisher
product
Washington's secret AI test is coming for open weights, and release dates go with it2 distinct publishers
security
The nationalization argument is really a vendor-continuity memo1 distinct publisher
build
MCP's roadmap fast-tracks five priorities and quietly queues everything else1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.
1 article · August 23, 2026