DeepSeek V4 Sparks a Pricing War Among China's AI Firms

DeepSeek's launch of its V4 models at aggressively low prices has set off a scramble across China's AI industry, sending shockwaves through developers, cloud providers, and infrastructure companies that suddenly found themselves racing to keep up. Weeks later, the aftershocks are still being felt.

Everyone is cutting prices fast

The speed of the response shows how competitive the market has become. Xiaomi, better known for smartphones and electric vehicles, slashed API costs for its MiMo-V2.5 model by 99%, a move that immediately resonated with developers. Usage exploded, with the model processing 1.7 trillion tokens in a single week, a jump of more than 999% from the week before.

AI unicorn MiniMax is taking a different approach with its newly launched flagship model, MiniMax M3, experimenting with a hybrid structure that pairs token-based billing with monthly subscription plans, an acknowledgment that not every user wants to pay the same way. The transition has not been entirely smooth. Some users found their token consumption running out far faster than expected, exhausting monthly quotas ahead of schedule. MiniMax apologised and moved to protect existing users with unlimited access, but the episode showed how delicate the balance between pricing innovation and user trust can be, a tension similar to what banks are managing as they build out AI infrastructure hubs like HSBC's in Singapore.

Beyond a simple price war

What is emerging in China's AI sector goes beyond a simple race to the bottom. Companies are layering in new subscription tiers, restructuring billing systems, and targeting different user groups with more tailored offerings. For enterprise customers running complex, high-volume workloads, the total cost of completing a task matters more than what a single token costs, and developers are rethinking pricing architecture around that insight, a shift that mirrors the broader move toward outcome-based pricing visible in how content teams now optimize for AI answer engines rather than search alone.

Cloud providers feel the squeeze

The pricing pressure is bleeding into the broader infrastructure that powers these companies. Tencent Cloud cut its API prices for DeepSeek's V4 models by as much as 97.5%, a reduction that would have been unthinkable a year ago.

The dynamic is structural. Because DeepSeek's models are open-weight, meaning multiple vendors can host them rather than a single provider controlling access, cloud companies are competing fiercely for the same developers. When anyone can offer the same model, price becomes the primary battleground. Research shows open models cost, on average, a fraction of what proprietary closed-source alternatives typically charge, and that gap is widening as competition intensifies. Chinese firms currently occupy the top positions globally for the most aggressive pricing discounts in AI inference.

The hidden stakes: data

Offering AI services cheaply is not only a competitive tactic. It also attracts users, and users generate the one resource that remains scarce: high-quality training data. Data availability continues to be one of the most significant constraints on improving AI model performance, which makes aggressive pricing look less like a sacrifice and more like a long-term investment.

The cost of cheap AI

Not everyone is celebrating falling prices. Companies that invested heavily in cloud computing infrastructure and server capacity to meet AI demand are now facing shifting economics. Shares in Chinese cloud and server-rental companies have softened since the price cuts began, reflecting investor concern about what a world of ultra-cheap AI inference means for businesses built around selling computing power.

China's AI pricing war is still very much in motion. AI services are getting cheaper faster than most anticipated, and the companies that survive this period will be the ones that build genuine value beyond the price tag. It is no longer only about who can build the most powerful model. It is about who can build a business around it.