Key Highlights:
- Chinese AI start up DeepSeek has made a bold move, the company just locked in a 75% price cut for its flagship V4-Pro AI model.
- The change is now permanent; DeepSeek made this official on May 23, 2026, making it clear they intend to reduce dependence on NVIDIA hardware following U.S. export restrictions.
Chinese artificial intelligence start up DeepSeek has permanently reduced prices for its flagship V4-Pro AI model by 75%, reflecting a major shift in China’s increasingly competitive AI industry. The company announced that API pricing for the model would now range between 0.025 yuan and 6 yuan per million tokens, compared with the previous pricing structure of 0.1 yuan to 24 yuan per million tokens.
By making the reduced prices permanent, DeepSeek is positioning itself more deeply in the Chinese AI market where companies are racing to attract developers and enterprise customers through lower costs and improved performance.
How DeepSeek’s Price Cut Highlights China’s Changing AI landscape
China's DeepSeek to make permanent 75% price cut on flagship V4‑Pro AI model https://t.co/p6VuFnq4Z7 https://t.co/p6VuFnq4Z7
— Reuters (@Reuters) May 23, 2026
DeepSeek’s announcement also reflects broader developments in China’s semiconductor and artificial intelligence ecosystem. According to the company, its V4 series models are optimized for Huawei’s Ascend 950 AI chips instead of relying heavily on NVIDIA processors which have faced tighter U.S. export controls in recent years.
DeepSeek did not directly state that improved access to Huawei chips enabled the lower pricing but the timing has drawn attention across the technology sector. China’s technology companies have been under pressure to develop domestic alternatives to foreign AI hardware, particularly after restrictions limited access to advanced NVIDIA chips widely used for training and operating large language models.
Making this price cut permanent is sure to escalate the price war already underway in China’s AI industry. Lower API costs will bring more developers and enterprise customers into the fold especially those on the search for affordable solutions in generative AI, automation, and smart enterprise software.
DeepSeek’s earlier models already drew attention by delivering strong results at significantly lower costs. Now, with this bold pricing strategy, DeepSeek is aiming to be both affordable and accessible, and to grab a bigger share of the Chinese AI market.
This shift also shows just how quickly costs are falling in the AI space. With improved chip efficiency and better access to domestic alternatives, the advanced AI services that used to seem out of reach are starting to look a lot more attainable for ordinary businesses and developers.
Also read: DeepSeek’s Two New V3.2 Models Are Immediate Rivals to GPT-5 & Gemini 3 Pro
Wrapping Up
DeepSeek’s decision to make a 75% price cut permanent isn’t just another marketing stunt, it’s a sign of how quickly the ground is shifting in China’s high-stakes AI race. By optimizing for Huawei’s Ascend 950 chips and reducing API prices, the company is betting hard on China’s ability to build a tough and self-reliant AI ecosystem.
The move is likely to increase pressure across the industry as rivals respond to falling prices and growing demand for affordable artificial intelligence services.
