China’s large artificial intelligence (AI) model market has witnessed a fresh wave of sweeping price reductions. Leading vendors have cut charges for Application Programming Interface (API) calls, computing power leasing and custom enterprise services. Combined with downward industry valuation revisions, the competitive paradigm for AI commercialization has been thoroughly reshaped. The sector has shifted from a race in technological iteration toward large-scale inclusive competition, representing the most prominent microeconomic change in today’s technology service industry.
This round of price cuts covers multiple subsegments, including general large models, industry-specific vertical models, and AI inference computing power. Market monitoring data indicates that several major players have reduced per-call prices for general text large models by 20% to 40%. Rates for high-frequency AI services such as image generation and code generation have dropped by over 30%. Barriers to lightweight AI packages for micro, small and medium-sized enterprises (MSMEs) have been substantially lowered, accompanied by a notable expansion of free trial quotas. In tandem with price adjustments, market valuations have retreated. Market capitalizations of frontrunners including Zhipu AI have recently fluctuated back to the HK$300 billion range, reflecting the capital market’s re-evaluation of profit models across the sector.
Structural shifts in supply and demand constitute the core driver behind these price cuts. Domestic capacity for large AI models has expanded rapidly over the past two years. Multiple technology firms have continued to invest heavily in computing infrastructure and model research and development, leading to a sharp surge in model supply and excess capacity for homogeneous general models. Meanwhile, demand growth from downstream consumers and enterprises has gradually slowed. Small merchants and traditional businesses tend to adopt AI in low-cost, lightweight scenarios rather than blindly pursuing top-tier model parameters. Their bargaining power has risen markedly, compelling upstream suppliers to seize market share through price cuts.
In terms of the competitive landscape among market participants, price cutting has become a shared strategy for smaller vendors seeking breakthroughs and leading firms aiming to consolidate moats. Leveraging strengths in lightweight models and niche scenarios, small and medium-sized AI enterprises adopt low-price tactics to tap into lower-tier markets such as e-commerce customer service, graphic and text creation, and digital transformation for small businesses, breaking the price monopoly held by major players. For their part, leading companies use large-scale price reductions to screen high-quality clients and phase out inefficient capacity. They also optimize model computing costs via technological upgrades, boosting market penetration through higher sales volume at lower prices and further squeezing the survival space for laggard enterprises.
These price adjustments have triggered ripple effects across upstream and downstream micro markets. For downstream MSMEs, the sharp drop in AI service costs has significantly lowered barriers to digital transformation. Small retailers, cultural and tourism operators, and local service merchants can affordably access AI-powered customer service, content marketing and data analytics tools, helping small businesses cut costs and improve efficiency. For the upstream industrial chain, pressure from price declines forces AI companies to accelerate business model transformation. Instead of merely selling model computing resources, they are shifting toward customized industry solutions and scenario-based value-added services to escape cutthroat low-price competition.
Industry analysts argue that the current wave of price cuts for large AI models does not amount to ruinous competition, but an inevitable outcome of market maturation. In earlier stages, the industry expanded rapidly driven by technological and capital dividends, accompanied by inflated valuations and redundant capacity. The ongoing price correction will accelerate industrial consolidation and phase out inefficient capacity with outdated technologies and weak real-world deployment capabilities. Going forward, competition will pivot from “price wars” toward a comprehensive contest over scenario implementation, technological iteration and service quality. Inclusiveness, refinement and vertical specialization will emerge as long-term trends for the large AI model industry.
As the peak season for corporate digital procurement arrives in the fourth quarter, industry insiders expect prices in the large AI model market to gradually stabilize. Enterprises with core technological barriers and mature application scenarios will be the first to restore profitability. The industry will officially move away from extensive expansion and enter a new phase of high-quality commercial deployment.