Model Quality Now Trumps Price as Demand Inflects
China's artificial intelligence market has entered a pivotal stage where model capability, not price, dictates market dominance. A March 27 JPMorgan report concludes that the industry is moving past the price wars of 2025 as demand for sophisticated coding and AI agent applications accelerates. In these complex, multi-step workflows, the value of reliability far outweighs the cost of tokens. The report illustrates this with a stark example: a model with a 98% success rate per step achieves a 67% final completion rate on a 20-step task, whereas a model with an 85% success rate collapses to just a 4% completion rate. This dynamic gives technically superior models significant pricing power.
The competitive landscape is defined by this performance gap. Chinese models have now achieved capabilities that surpass leading U.S. models from a year prior, while offering more economical pricing for the local market. However, with customer focus locked on task completion, the cheapest models often result in the highest effective cost. As a result, companies with leading-edge models can expand into lower-end markets, but firms competing only on price will find it nearly impossible to move upmarket.
JPMorgan Forecasts 2029 Profitability for Sector Leaders
JPMorgan projects that the industry's intense competition will forge a few key winners, forecasting that leaders Zhipu and MiniMax will reach profitability starting in 2029. The bank reiterated its "Overweight" ratings for both companies, setting price targets of HK$800 for Zhipu and HK$1100 for MiniMax. This outlook is supported by a potential for explosive growth mirroring the U.S. market, where Anthropic's annual recurring revenue (ARR) grew approximately 19-fold in 15 months, from $1 billion in December 2024 to $19 billion by March 2026.
This growth trajectory aligns with China's national strategy, which includes an "AI plus consumption" initiative and aims to build an AI-related sector worth over 10 trillion yuan by 2030. As major tech firms like Tencent and Alibaba integrate advanced AI tools into their ecosystems, the market is shifting from developer experiments to full-scale enterprise deployment. This transition underscores the immense pressure on AI firms to continuously innovate, as profitability hinges on gross profit growth outpacing the relentless spending required for research and development.
