Indian Companies Shift to Chinese AI Models to Cut Costs

Indian firms are increasingly adopting large language models (LLMs) from Chinese providers such as DeepSeek, Alibaba and Moonshot AI to reduce artificial‑intelligence spending. The move follows a sharp rise in token prices for U.S. models; for example, OpenAI’s GPT‑5.5 series costs $5‑$12 per million input tokens and $30‑$54 per million output tokens, whereas DeepSeek models accessed via Microsoft’s Foundry platform in southern India charge 19 cents‑$1.74 per million input tokens and 51 cents‑$5.40 per million output tokens. “The US models are expensive, and for a lot of basic things, you don’t need them,” said Gupta, adding that they are “overkill, like trying to drive a sports car on a crowded city road.”

The adoption coincides with OpenAI and Anthropic expanding their Indian presence, but cost pressures are driving many startups and large enterprises toward the lower‑priced Chinese alternatives. Usage of Chinese LLMs in India more than doubled to 25 trillion tokens in the last week of June, representing a 78 % increase over U.S. models. Global companies such as Tesla, Amazon, Uber and Walmart have also begun capping AI usage and turning to Chinese models for cost efficiency.