Breaking News
DeepSeek said it will permanently maintain a 75% price cut on its flagship V4-Pro artificial intelligence model, a move that could intensify pricing pressure across China’s rapidly expanding AI market.
The Chinese startup said on Saturday that API pricing for V4-Pro would remain at between 0.025 yuan and 6 yuan per million tokens, down from previous rates of 0.1 yuan to 24 yuan depending on usage type.
The decision highlights how Chinese AI firms are increasingly competing not only on model performance but also on cost efficiency and infrastructure optimization as they race to scale adoption across enterprises and developers.
DeepSeek did not specify whether the permanent price reduction was linked to improved availability of Huawei’s Ascend 950 AI chips, which the company has been using to maximize the performance of its V4 models.
That question matters because compute availability has emerged as one of the defining competitive constraints in China’s AI sector following U.S. export controls that restrict Chinese companies’ access to the most advanced AI chips from Nvidia.
Huawei’s Ascend processors have increasingly become a domestic alternative powering China’s AI ecosystem, although separate restrictions on semiconductor manufacturing equipment continue to limit Huawei’s ability to rapidly scale production capacity.
When DeepSeek launched V4 last month, it had warned that the more powerful Pro version would cost substantially more than its Flash variant because of shortages in high-end compute infrastructure.
At the time, the company said prices could decline sharply once Huawei’s Ascend 950 supernodes entered larger-scale production later this year.
The permanent pricing shift now suggests China’s AI infrastructure ecosystem may be stabilizing faster than expected — or that DeepSeek is willing to aggressively compress margins to capture developer and enterprise market share.
The move also reinforces how AI competition is increasingly evolving into an infrastructure and economics battle rather than purely a model capability race.
DeepSeek shocked the global technology industry earlier this year after releasing AI systems that rivaled leading Western models at significantly lower operating costs, fueling concerns among U.S. technology firms and investors about the sustainability of current AI spending levels.
Its latest pricing decision could place further pressure on both Chinese and international AI providers already grappling with rising compute costs, expensive GPU infrastructure, and intensifying competition around enterprise AI adoption.
The broader implication is that China’s AI market may increasingly prioritize affordable, high-volume AI deployment powered by domestic semiconductor ecosystems rather than competing solely on frontier-model prestige.
The Chinese startup said on Saturday that API pricing for V4-Pro would remain at between 0.025 yuan and 6 yuan per million tokens, down from previous rates of 0.1 yuan to 24 yuan depending on usage type.
The decision highlights how Chinese AI firms are increasingly competing not only on model performance but also on cost efficiency and infrastructure optimization as they race to scale adoption across enterprises and developers.
DeepSeek did not specify whether the permanent price reduction was linked to improved availability of Huawei’s Ascend 950 AI chips, which the company has been using to maximize the performance of its V4 models.
That question matters because compute availability has emerged as one of the defining competitive constraints in China’s AI sector following U.S. export controls that restrict Chinese companies’ access to the most advanced AI chips from Nvidia.
Huawei’s Ascend processors have increasingly become a domestic alternative powering China’s AI ecosystem, although separate restrictions on semiconductor manufacturing equipment continue to limit Huawei’s ability to rapidly scale production capacity.
When DeepSeek launched V4 last month, it had warned that the more powerful Pro version would cost substantially more than its Flash variant because of shortages in high-end compute infrastructure.
At the time, the company said prices could decline sharply once Huawei’s Ascend 950 supernodes entered larger-scale production later this year.
The permanent pricing shift now suggests China’s AI infrastructure ecosystem may be stabilizing faster than expected — or that DeepSeek is willing to aggressively compress margins to capture developer and enterprise market share.
The move also reinforces how AI competition is increasingly evolving into an infrastructure and economics battle rather than purely a model capability race.
DeepSeek shocked the global technology industry earlier this year after releasing AI systems that rivaled leading Western models at significantly lower operating costs, fueling concerns among U.S. technology firms and investors about the sustainability of current AI spending levels.
Its latest pricing decision could place further pressure on both Chinese and international AI providers already grappling with rising compute costs, expensive GPU infrastructure, and intensifying competition around enterprise AI adoption.
The broader implication is that China’s AI market may increasingly prioritize affordable, high-volume AI deployment powered by domestic semiconductor ecosystems rather than competing solely on frontier-model prestige.
See What’s Next in Tech With the Fast Forward Newsletter
Tweets From @varindiamag
Nothing to see here - yet
When they Tweet, their Tweets will show up here.
