How much longer can China afford cheap AI?
China is experiencing a rapid proliferation of large language models (LLMs), with nearly 1,000 models now available in the market. This surge is driven by significant government and private sector investment, aiming to establish China as a leader in AI. However, the cost of developing and maintaining these advanced models is substantial, raising questions about the long-term affordability and sustainability of this growth. The intense competition and the need for high-end computing resources are putting pressure on companies, potentially leading to consolidation or increased government subsidies to sustain the current pace of innovation.
The sheer volume of nearly 1,000 large language models in China suggests a vibrant, if perhaps overheated, market. This rapid expansion is a direct outcome of Beijing's strategic push for AI self-sufficiency and leadership, with significant capital flowing into AI development. However, the underlying economics of this boom are challenging. Training and operating these models demand immense computational power and specialized chips, which remain expensive, particularly given export controls on advanced semiconductors. Our view is that the current pace of LLM development in China is not sustainable without significant government intervention or a major shift in chip supply. The cost pressure will likely lead to consolidation, favoring well-capitalized tech giants like Baidu or Alibaba, or those with strong state backing. For other Asian markets, this dynamic in China points to the critical importance of secure and affordable access to high-performance computing infrastructure for their own AI ambitions. The cost of AI innovation is a regional challenge, not just a Chinese one.
Related reading
6 stories
Semiconductor Investing Gets More Selective as HBM, CPU and NAND ‘Pinpoint’ ETFs Flood the Market
The high cost of advanced semiconductors is central to China's AI challenge; we recently covered chip investing.
People’s Daily rejects US claims of malicious AI distillation, warns of countermeasures
Databricks to Invest Over US$350 Million in Singapore, Double Its Workforce
Personetics Launches AI Banking Console for Relationship Managers
Anthropic Sets October Launch for Singapore Office With Local Hiring Planned
