The record
Written from the 1 report below. Nothing here is unsourced.
- Chinese AI models such as DeepSeek, Qwen, GLM and MiniMax are seeing rapid global adoption because they cost 20 to 170 times less than Western proprietary models while nearing their performance.
- On the OpenRouter platform, Chinese models recently processed 4.1 trillion tokens in a week, overtaking US models for the first time.
- Indian companies are quietly evaluating these models, with Krutrim having deployed DeepSeek R1 on Nvidia H100 GPUs hosted in India.
- However, no large listed Indian company has disclosed buying Chinese AI tokens directly, as firms serving US and European customers worry about data governance and regulatory compliance.
What to watch next
- Whether more Indian enterprises publicly disclose use of Chinese AI models or follow Krutrim's model of hosting them on Indian infrastructure
- How data governance and regulatory compliance concerns shape enterprise deployment choices, especially for firms serving US and European clients
- Whether cheap imported models discourage investment in India's own foundation models
Coverage1
1 report
English national1
All filed from India
Named China · India · United States · Anthropic · OpenAI · Alibaba · Bhavish Agrawal · Brian Armstrong · Coinbase · DeepSeek R1 · GLM 5.2 · Kimi 2.7 · Krutrim · MiniMax · Moonshot AI
The 1 report is listed beside the record.
Ask this story
Answers cite the reports above, or say they can't.