Skip to main content
GPUBeat Frontier Models DeepSeek Cuts AI Model Prices by…

DeepSeek Cuts AI Model Prices by 75% Amid China’s Chip Strategy Shift

DeepSeek has slashed its V4-Pro model prices by 75%, intensifying the AI pricing war and reflecting shifts in China's semiconductor landscape.

Chinese AI startup DeepSeek has made waves in the artificial intelligence sector by permanently reducing the prices of its flagship V4-Pro model by an astounding 75%. This aggressive move is a escalation in the ongoing pricing war among AI firms in China, aimed at challenging the dominance of U.S. competitors. The price cut comes as domestic companies navigate the complexities posed by U.S. semiconductor restrictions.

The new pricing for DeepSeek’s V4-Pro application programming interface now ranges from 0.025 yuan to 6 yuan per million tokens, a sharp drop from the previous range of 0.1 yuan to 24 yuan. In practical terms, accessing this advanced large language model now costs less than $0.85 per million tokens, down from approximately $3.30. This pricing strategy positions DeepSeek well below many Western models, emphasizing a clear shift towards affordability and scalability in the AI marketplace.

DeepSeek's announcement indicates that the company may be overcoming previous limitations tied to compute capacity, a challenge exacerbated by U.S. export controls on high-end chips. While the firm did not directly link the price cut to supply improvements, there are signs that Huawei’s Ascend 950 AI processors are becoming more readily available as they approach large-scale production. When DeepSeek launched the V4-Pro last month, it noted that the model's higher costs were largely due to constraints in high-end compute capabilities, with prices noted to be up to 12 times higher than its Flash model.

The implications of DeepSeek's price reduction extend beyond mere numbers. It reflects broader dynamics within China's semiconductor supply chain, where local firms are rapidly innovating to reduce reliance on American chipmakers like Nvidia. The tightening of U.S. export restrictions has created a vacuum, prompting domestic companies to forge ahead with their own solutions. Huawei, in particular, has emerged as a key player, with its Ascend AI chips being increasingly adopted by Chinese cloud providers and AI startups.

See also  Nyle DiMarco Inspires with Commencement Speech at Gallaudet University

However, China’s ambitions in the chip sector face challenges. U.S.-led restrictions on semiconductor manufacturing equipment continue to hinder Huawei’s ability to scale up production of advanced processors, particularly at the advanced nodes necessary to compete with Nvidia's latest offerings. This complex environment has led to a bifurcated AI market in China, characterized by rapid software advancements alongside ongoing hardware limitations.

As the space evolves, the stakes are high for both Chinese and American companies. DeepSeek's significant price slashes highlight the urgency within China’s AI sector to innovate and capture market share. With the domestic industry quickly adapting to shifting dynamics, the effects of this pricing strategy will likely resonate not only across Asia but also globally, prompting traditional players to reassess their positions in an increasingly competitive environment.

GD

GPUBeat Desk

Desk · joined 2026

GPUBeat Desk covers AI infrastructure — chips, foundation models, inference economics, datacenter buildouts, and the geopolitics of compute.