---
title: "Microsoft books Azure revenue on Chinese open models running on Baseten hardware it does not own"
description: "Microsoft now bills customers' committed Azure dollars for open models from Chinese labs running on Baseten's hardware, a toll booth earning rent on compute it does not own."
publisher: "The Inference"
section: "Model"
published: 2026-11-18T16:27:49.724Z
modified: 2026-11-18T16:27:49.724Z
canonical: https://theinference.org/article/microsoft-books-azure-revenue-on-chinese-open-models-running-on-baseten-hardware-it-does-not-own
language: en
keywords: "Microsoft, Baseten, Azure, AI Models, Inference, Open Models"
---

# Microsoft books Azure revenue on Chinese open models running on Baseten hardware it does not own

> Microsoft now bills customers' committed Azure dollars for open models from Chinese labs running on Baseten's hardware, a toll booth earning rent on compute it does not own.

[Microsoft](https://theinference.org/markets/companies/microsoft)'s model catalog now sells compute Microsoft does not own. On November 17, 2026, [Baseten](https://theinference.org/markets/companies/baseten), a San Francisco inference company founded in 2019, valued at $13 billion after raising more than $2 billion, [went live in public preview inside Microsoft Foundry](https://www.baseten.co/blog/baseten-is-now-available-on-microsoft-foundry/), the Azure platform where enterprises build and govern AI applications [1][2]. Four frontier open models, [DeepSeek](https://theinference.org/markets/companies/deepseek) V4.1 Flash, GLM-5.3, GLM-5.3-Flash and Kimi K3, all run on Baseten's infrastructure, not Azure's own GPUs [1][3].

## Every dollar of it counts against Azure commitments

The models are sold as first-party Foundry consumption, and 100% of the spend counts toward a customer's Microsoft Azure Consumption Commitment, the pre-purchased cloud dollars enterprises already owe [1][3]. Microsoft books Azure revenue on hardware it did not buy, paid for out of contracts signed for its own capacity. Administrators opt in, deployments go live in seconds, requests are stateless with zero data retention, and general availability is promised for early 2027 [1][2].

*Chart: **Microsoft's own capex has nearly tripled since mid-2024** Line chart of Microsoft's quarterly capital spending and operating cash flow, in billions of dollars, from the quarter ended June 2024 to the quarter ended September 2026: capex climbs from $13.9 billion to $39.4 billion while operating cash flow rises from $37.2 billion to $58.0 billion.*

*Microsoft's capital spending and operating cash flow by quarter, in billions of dollars, from its SEC filings. Capex rose from $13.9 billion in the quarter ended June 2024 to $39.4 billion in the quarter ended September 2026, up 103% year over year, while operating cash flow rose from $37.2 billion to $58.0 billion. Source: Microsoft quarterly filings. [9]*

## Microsoft's own coding model wins on a price that expires in May

The launch lineup leans on Chinese labs: DeepSeek, Zhipu's GLM and [Moonshot](https://theinference.org/markets/companies/moonshot)'s Kimi. The same week, Microsoft shipped MAI Code 2, pitching 96% of Claude Opus 5.5's performance at 5% of the cost, but that math rests on a 50% promotional price that expires in May 2027 [6][7]. Independent of promotion, [a community evaluation on Hacker News](https://news.ycombinator.com/item?id=50539594) put DeepSeek V4.1 at 77.0 on DeepSWE 1.1 against 73.2 for MAI-2, a figure [InfoWorld's report](https://www.infoworld.com/article/4250152/microsoft-launches-mai-code-2-models-to-challenge-openai-and-anthropic-in-ai-coding.html) of Microsoft's own benchmark tables matches [5][7]. The comparison is uncontrolled, as its author says.

*Chart: **A coding task on MAI Code 2 costs $0.31 only at the 50% promo price** Bars comparing average cost per coding task: MAI-2 at $0.31 under promotional pricing, with a whisker to its $0.62 list price after the promotion ends, GPT-6.1 Sol at $0.76 and Claude Opus 5.5 at $6.31.*

*Average cost per coding task across four coding benchmarks, in dollars, as reported for Microsoft's MAI Code 2 launch. The MAI-2 bar is an estimate at 50% promotional pricing, which expires in May 2027; the whisker shows the roughly $0.62 a task the same comparison implies at list price. Sources: Brave New Coin, InfoWorld. [6][7]*

## Baseten gets the rails, Microsoft carries the question

Baseten gets Microsoft's distribution and billing rails, on top of annualized revenue [the Wall Street Journal reported at around $600 million](https://www.pymnts.com/news/investment-tracker/2026/baseten-nears-1-5-billion-funding-round-as-inference-demand-surges/) in the spring [8]. The Chinese labs get first-party placement in the world's largest enterprise cloud. Enterprises get cheaper frontier intelligence on an invoice that credits their commitments. Microsoft shareholders hold the open question: whether a pass-through with no disclosed margin split earns more than [renting Azure's own GPUs](https://theinference.org/article/microsoft-borrows-25-billion-to-own-the-ai-capacity-it-used-to-rent). Watch the first Azure bills under this arrangement, and whether MAI's discounted pricing holds after May 2027 [1][7].

## Takeaway

Baseten's inference service now lives inside Microsoft Foundry, and every dollar customers spend on its Chinese open models counts toward their prepaid Azure commitments, so Microsoft books cloud revenue on GPUs it never bought, at a margin split nobody has disclosed.

## Sources

1. [Baseten blog, Baseten is now available on Microsoft Foundry, 17 November 2026](https://www.baseten.co/blog/baseten-is-now-available-on-microsoft-foundry/)
2. [Yahoo Finance / GlobeNewswire, Baseten Brings High-Performance Inference for Open Models to Microsoft Foundry, 17 November 2026](https://finance.yahoo.com/technology/ai/articles/baseten-brings-high-performance-inference-231500946.html)
3. [Unite.AI, Baseten Brings Four Open Models to Microsoft Foundry Preview, 17 November 2026](https://www.unite.ai/baseten-brings-four-open-models-to-microsoft-foundry-preview/)
5. [Hacker News, spwa4 comment on MAI Code 2 and Code 2 Flash, November 2026](https://news.ycombinator.com/item?id=50539594)
6. [Brave New Coin, Microsoft Takes Aim at OpenAI and Anthropic With Cheaper AI, But Read the Fine Print, 18 November 2026](https://bravenewcoin.com/insights/microsoft-mai-code-2-ai-price-war)
7. [InfoWorld, Microsoft launches MAI Code 2 models to challenge OpenAI and Anthropic in AI coding, 18 November 2026](https://www.infoworld.com/article/4250152/microsoft-launches-mai-code-2-models-to-challenge-openai-and-anthropic-in-ai-coding.html)
8. [PYMNTS, Baseten Nears $1.5 Billion Funding Round as Inference Demand Surges, 18 June 2026](https://www.pymnts.com/news/investment-tracker/2026/baseten-nears-1-5-billion-funding-round-as-inference-demand-surges/)
9. [Microsoft, quarterly figures from its SEC filings, retrieved 18 November 2026](https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000789019)


---

Microsoft books Azure revenue on Chinese open models running on Baseten hardware it does not own — The Inference. Canonical: https://theinference.org/article/microsoft-books-azure-revenue-on-chinese-open-models-running-on-baseten-hardware-it-does-not-own
