---
title: "Enterprise AI hit a US$177bn run rate, Menlo estimates; Anthropic and OpenAI take 86%"
summary: "Menlo Ventures’ estimates of enterprise AI spending, 2023 to September 2026, and of model API shares, checked against Anthropic’s own filing. Menlo is an Anthropic investor."
url: https://www.thetechbag.com/blog/enterprise-ai-spend-177bn-anthropic-openai-86-percent
format: Trend
desk: "AI in the enterprise"
region: Global
published: 2026-11-04
facts_checked: 2026-11-04
author: TechBag (The TechBag Journal)
tags: ["enterprise AI spending","Menlo Ventures","Anthropic","OpenAI","Google Gemini","LLM API market share","AI agents","AI coding","Claude Code","Codex","open-weight models","token prices","Anthropic S-1"]
cover: https://www.thetechbag.com/blog/covers/enterprise-ai-spend-177bn-anthropic-openai-86-percent.png
---

# Enterprise AI hit a US$177bn run rate, Menlo estimates; Anthropic and OpenAI take 86%

*Menlo Ventures’ estimates of enterprise AI spending, 2023 to September 2026, and of model API shares, checked against Anthropic’s own filing. Menlo is an Anthropic investor.*

Trend · AI in the enterprise · By TechBag · Published 2026-11-04 · Facts checked 2026-11-04 · Canonical: https://www.thetechbag.com/blog/enterprise-ai-spend-177bn-anthropic-openai-86-percent

The short version

1. Menlo Ventures estimates that enterprise spending on putting AI to work reached an annualised US$176.8bn by September 2026, 5.4 times its restated US$33bn for 2025 and about 100 times its US$1.7bn for 2023.[1]
2. In its survey, Anthropic’s share of enterprise model-API spending rose from 40% to 55% and OpenAI’s from 27% to 31%: 86% together, up from 67%. Google’s fell from 21% to 9%. Menlo is an investor in Anthropic.[1][2]
3. A filing points the same way: Anthropic’s registration statement reports US$17.0bn of revenue in the quarter to 30 September, against US$1.4bn a year earlier.[3]

**Who published this, and how.** Menlo Ventures, a US venture firm, published the fourth edition of its enterprise AI report on 3 November. It marks Anthropic as a “Menlo Ventures portfolio company”, as it does OpenRouter, Fireworks and Modal, which the report also cites.[1] Its survey covered nearly 400 US enterprise decision-makers responsible for more than US$4bn of AI spending, fielded by Morning Consult from 8 to 26 September 2026. The dollar figures are Menlo’s own estimates from a bottom-up model “informed by survey responses, public filings, and third-party reports”, shown as annualised run rates.[1] This piece sets out what the report measures, where it says the money goes, how one company’s filing compares, and what the numbers cannot tell you.

## The line: US$1.7bn to US$177bn in three years

Menlo splits AI spending in two. **Deployment**, using AI, was US$176.8bn: US$99.5bn of infrastructure (model inference sold by the labs, inference platforms and agent infrastructure) and US$77.3bn of applications. **Development**, building models, was a further US$102.5bn: US$87.2bn of training compute and US$15.3bn of training data and services. The headline number is deployment only.[1]

**Figure: Enterprise spending on using AI grew about 100-fold in three years, by Menlo's estimate**

*Enterprise generative AI deployment spend, US$bn, annualised run rate, 2023 to September 2026, Menlo's restated series*

|  | Deployment spend |
| --- | --- |
| 2023 | 1.7 |
| 2024 | 8.5 |
| 2025 | 33 |
| Sep 2026 | 176.8 |

Source: Menlo Ventures, 2026: The State of Enterprise AI · as of 3 November 2026 · Menlo estimates, not reported revenue. Prior years reclassified to the 2026 taxonomy

ReportedWhy last year's US$37bn is not on this chart

Menlo’s December 2025 report put 2025 spending at US$37bn and 2024 at US$11.5bn, on a scope that excluded “inference and model serving” sold by the clouds and inference platforms.[2] The 2026 report restates the earlier years to its new taxonomy, which moves data-labelling vendors such as Mercor from deployment to development: 2024 becomes US$8.5bn and 2025 US$33bn.[1] Set against the old US$37bn, the new US$176.8bn would mix two definitions, so we chart only the restated series.

Menlo’s explanation for the jump is agents: software that carries out multi-step work rather than answering one question. Spending averaged 4.7 times the previous year’s against a planned 1.9 times; the median company spent 2.2 times against a planned 1.8 times. The share of enterprises using AI rose from 55% to 65%, so most of the growth came from more spending per company, not more companies.[1]

## Where it goes: two labs, and coding

Menlo’s shares are “estimated dollars spent based on proportion of production API usage”: respondents reported which models ran their AI workloads, weighted by their AI spending.[1] On that measure Anthropic went from 40% to 55%, which Menlo attributes to its Opus models in long-running tasks and coding; OpenAI from 27% to 31%; Google from 21% to 9%, its spending now skewed to cheaper Flash models.[1][2]

**Figure: Anthropic and OpenAI went from 67% to 86% of enterprise API spend; Google's share fell by more than half**

*Share of enterprise LLM API spend by model provider, %, Menlo's 2025 and 2026 reports*

|  | 2025 report | 2026 report | Change |
| --- | --- | --- | --- |
| Anthropic | 40 | 55 | 15 pts |
| OpenAI | 27 | 31 | 4 pts |
| Google | 21 | 9 | -12 pts |
| Everyone else | 12 | 5 | -7 pts |

Source: Menlo Ventures, State of Generative AI in the Enterprise (9 December 2025) and State of Enterprise AI (3 November 2026) · as of 3 November 2026 · Everyone else = 100% minus the three, TechBag arithmetic. Survey-weighted estimates; Menlo is an Anthropic investor

**Applications.** Coding was US$34.0bn, up 7.7 times, and the labs’ own tools, Claude Code and Codex, took US$24.0bn of it (70%). Enterprise chatbots were US$25.0bn. The two make up 76% of the US$77.3bn application market; healthcare, third, was US$3.7bn.[1]

**Open models.** Open-weight models handled 19% of tokens but took about 4% of model spending, because their average price fell 3.3 times to US$0.13 per million tokens, against a blended US$0.70 for closed models, down 24%.[1]

ReportedA check from a filing: Anthropic's own revenue

Anthropic’s amended S-1, filed on 2 November for a share listing, reports quarterly revenue of US$1,382m for July to September 2025, US$2,040m, US$4,731m, US$11,555m, and US$16,956m for July to September 2026. In the first nine months of 2026, 79% of revenue was consumption-based (usage) and 21% subscriptions; US$23.47bn of the US$33.24bn was billed to US addresses. It says price cuts with new models and contract discounts took about US$600m off revenue between the June and September quarters.[3] These are global revenues, consumers included, so they do not test Menlo’s 55% share; they confirm the scale of the growth. OpenAI publishes no comparable filing.

## What the numbers cannot tell you, and what to watch

**The limits.** The survey is of US companies. The spending totals are modelled estimates, and annualised run rates are not a year’s revenue. Menlo restates its own past figures, so editions compare only on the newest basis. Some comparisons mix groups: agents are used by 74% of AI builders this year, against 27% of startup deployments in the 2025 survey. And the publisher holds stakes in the leading lab and in several infrastructure firms the report names.[1][2]

Our viewOur view: what concentration means for a buyer

When two suppliers carry 86% of API spending, the terms matter as much as the model. Anthropic’s filing says a growing number of enterprise customers sign non-cancelable annual or multi-year contracts with minimum spend commitments in exchange for a committed price list.[3] A buyer weighing such a commitment can ask what happens to the price list when new models arrive, and how much work could move to another model without a rewrite.

**What to watch.** Anthropic’s run-rate figure for 31 October is left blank in its 2 November amendment, so the number the company itself uses for its current pace is not yet public.[3] Menlo’s next edition will show whether its 86% holds after the discounts on flagship models it says began “in recent weeks”.[1] On the return side, our [August analysis of McKinsey’s survey](https://www.thetechbag.com/blog/ai-return-few-firms-see-profit) found 37% of firms reporting any profit from AI.[4]

India: neither document gives an Indian figure

Menlo’s survey covers US enterprises only. Anthropic’s filing splits revenue into the United States and “International”, which was US$9.77bn, or 29%, of the nine months; India appears only in the offering’s selling restrictions.[1][3] We found no published Indian estimate built on a stated method to set beside these.

How we did this

Figures as of 3 November 2026

Every spend, share, price and survey figure is Menlo Ventures’, read from its 3 November 2026 report and, for the 2025 shares and the old US$37bn, its 9 December 2025 report. The chart uses Menlo’s restated 2023–2025 values. “Everyone else” in the share chart is 100% minus the three named providers (TechBag arithmetic; Menlo says near-frontier and open models held “just under 5%”). Anthropic’s revenue figures are from its Form S-1/A filed on 2 November 2026; the US share of its nine-month revenue is TechBag arithmetic. We did not use Menlo’s estimate of the labs’ share of cloud revenue, which it does not source, or any figure published after 7pm IST on 4 November.

Spotted an error? Write to us; corrections are published at the foot of the piece.

Sources

1. [Menlo Ventures, 2026: The State of Enterprise AI (incl. Methodology)](https://menlovc.com/perspective/2026-the-state-of-enterprise-ai/) — 3 November 2026 Marks Anthropic, OpenRouter, Fireworks and Modal as Menlo Ventures portfolio companies.
2. [Menlo Ventures, 2025: The State of Generative AI in the Enterprise](https://menlovc.com/perspective/2025-the-state-of-generative-ai-in-the-enterprise/) — 9 December 2025
3. [Anthropic, PBC, Form S-1/A (Amendment No. 1), registration statement](https://www.sec.gov/Archives/edgar/data/2130357/000162828026068676/anthropicpbc-sx1ax1.htm) — 2 November 2026 Quarterly results of operations, Note 2 (revenue by region and type), and the Run-Rate Revenue definition.
4. [The TechBag Journal: 37% of firms see any profit from AI, 6% a real one, as agents spread](https://www.thetechbag.com/blog/ai-return-few-firms-see-profit) — 26 August 2026

**Disclosure.** TechBag is a software reseller: it earns a margin when a reader buys through it, including from vendors named here. The analysis is TechBag’s own; no vendor paid for or saw this piece before publication. Facts are labelled reported, TechBag count or our view.

**Read this next:** [37% of firms see any profit from AI, 6% a real one, as agents spread](https://www.thetechbag.com/blog/ai-return-few-firms-see-profit) — Where the spending meets returns: 37% of firms see any profit from AI and 6% a real one, in McKinsey’s survey of 1,719 executives.

Go deeper on TechBag

- [Developer tools guide](https://www.thetechbag.com/browse/devops/developer-tools) — where AI coding assistants, the largest app market in the report, sit in a toolchain
- [Microsoft Copilot](https://www.thetechbag.com/microsoft/microsoft-copilot) — the workplace assistant Menlo names in its US$25bn chatbot market
- [Databricks Agent Bricks](https://www.thetechbag.com/databricks/databricks-agent-bricks) — agent infrastructure that runs on Claude, OpenAI and other models