AI Read the original on Artificialintelligence-news 2 min read 1

Microsoft becomes primary supplier of OpenAI models to China

Microsoft has emerged as the primary gateway for OpenAI's large language models within the Chinese market. While OpenAI and Anthropic refuse to sell directly to China due to intellectual property and misuse risks, Microsoft leverages its unique contract to supply these technologies to major internet giants. This arrangement allows the tech giant to capture significant revenue from Chinese firms like ByteDance while navigating complex geopolitical tensions and regulatory scrutiny.

Скляна іконка з білим геометричним логотипом OpenAI, що висить у центрі перед однотонним сірим фоном із м'яким світлом знизу.
Скляна іконка з білим геометричним логотипом OpenAI, що висить у центрі перед однотонним сірим фоном із м'яким світлом знизу. · Image source: Artificialintelligence-news

According to Artificialintelligence-news, reporting originally by Bloomberg, Microsoft has quietly secured a dominant position as the main supplier of OpenAI models to China's largest internet companies. This development places Microsoft in a unique situation where it serves as the sole American AI vendor providing GPT series technology to Chinese firms that the model's own creator refuses to engage with directly.

Significant scale and key corporate clients

The commercial scope of this arrangement is substantial, involving some of China's most influential technology players. ByteDance has reportedly become Microsoft’s largest AI customer in recent years, utilizing OpenAI models extensively. The company is currently on track to spend more than $1 billion annually on Microsoft’s cloud and AI services.

Other major entities are also participating in this ecosystem through the Azure platform:

  • Ant Group: Purchases models via Azure while simultaneously developing internal systems.
  • Meituan: Utilizes AI models provided through Microsoft's infrastructure.
  • Tencent: Integrates AI capabilities into its services via the Azure partnership.

Internal data suggests this growth is a major success for Microsoft. During a July 2025 sales meeting, then-chief commercial officer Judson Althoff informed staff that Azure’s AI revenue in China expanded faster than any other territory, roughly tripling in the financial year to June 2025 after a 400% increase the previous year.

Navigating risk and geopolitical tensions

Microsoft's ability to facilitate these sales stems from its singular contract with OpenAI, which grants it the authority to set terms for international distribution. Because Anthropic has opted out of the Chinese market entirely, Microsoft acts as a necessary intermediary for models whose developers view China as too high-risk. To mitigate concerns regarding model "distillation"—where one model's output is used to train another—Microsoft employs automated monitoring and restricts sales to established corporations rather than individual developers.

To further manage risk, Microsoft does not host OpenAI models on Chinese soil; instead, customers access them via the internet from international data centers, including those in Singapore. This creates a complex dynamic where Microsoft is simultaneously selling American models into China and hosting Chinese models, such as DeepSeek, for Western enterprise clients. While this strategy maximizes profit margins, it faces ongoing scrutiny from U.S. lawmakers who view China's AI advancements as a strategic threat.

The company currently balances these conflicting interests by maintaining a delicate technological bridge between the United States and China. This position ensures Microsoft remains the only major player being paid by both sides of the global AI trade while navigating intense political pressure.

FAQ

How does Microsoft provide OpenAI models to Chinese companies?
Microsoft uses its unique contract with OpenAI to act as an intermediary. Because OpenAI and Anthropic refuse direct sales to China, Microsoft provides access via the Azure platform. Customers access these models over the internet from international data centers, such as those in Singapore, rather than on Chinese soil.
What risks does Microsoft manage when selling AI to China?
Microsoft manages risks by employing automated monitoring and restricting sales to established corporations instead of individual developers. This strategy aims to prevent model distillation, where one model's output is used to train another, while navigating geopolitical tensions and U.S. lawmaker scrutiny.
Telegram

Fresh news on our Telegram

Get instant alerts for new posts in «AI»

@proaiandevenmore