TMTPOST--Since the release of ChatGPT on November 30, 2022, nearly 80,000 artificial intelligence (AI)-related companies in China have either been dissolved, revoked, or shut down. This number represents about 8.9% of the 878,000 newly registered businesses in this period, according to data compiled by TMTPost AGI as of Tuesday.

In the past three years alone, over 200,000 AI companies in China have faced closure, while a total of 353,000 AI-related enterprises have vanished over the last decade due to various reasons including liquidation, suspension, and deregistration. As of August 7, 2024, the country saw the establishment of 300,700 new AI-related companies this year.

Currently, there are 1.804 million AI-related enterprises in operation in China. Data from the Ministry of Industry and Information Technology shows that over 4,500 companies are in the AI industry system. Additionally, more than 180 generative AI models registered with the authorities are being actively used, with a registered user base of over 564 million.

Despite the initial surge in AI startups fueled by the technology boom, many companies have struggled with financial viability, facing the dual challenges of high cash burn rates and a cooling investment climate.

The situation in China is not alone. Earlier this month, AI unicorn Character.AI was acquired by Google for $2.5 billion, with its founders joining Google's AI team. Other acquisitions include Microsoft’s purchase of Inflection, and Amazon’s acquisition of Adept. Even in the U.S., AI startups are feeling the pinch, with seed-stage funding plunging 70% in the first quarter of 2024 compared to the previous year.

China has also witnessed a slowdown in AI investment. According to CB Insights, a total of 232 investment deals in AI sector was closed in 2023, a 38% decrease from the previous year, with the total amount of funding plummeting by 70%. The Cleantech Research Center reports a continued decline in equity investment activity, with significant reductions in early-stage investments and venture capital funding.

Ren Zeping, Vice President of the China Private Economy Research Association, highlights that the average lifespan of Chinese small and medium-sized enterprises (SMEs) is around three to five years, with smaller businesses averaging just 2.5 years. This short lifespan reflects the high failure rate of new ventures, particularly in the AI sector.

AI industry leaders, including Wang Xiaochuan, the CEO of Baichuan AI, predict a “survival of the fittest” scenario, with only a handful of major players expected to remain relevant. Wang emphasized that while the sector will continue to grow, it will also experience significant consolidation.

Robin Li, the founder and CEO of Baidu, said that many large models are essentially "reinventing the wheel," and the "hundred-model battle" is wasting resources. He noted that over 200 domestic models have minimal usage, and no single company can fully capitalize on all the opportunities in the era of large models.

"Competition is normal. When new things and big opportunities arise, everyone rushes in, and then a process of ‘sifting through the sand’ follows," Li remarked.

Regarding the gap between large models in China and the U.S., Li emphasized that there is not much difference in the technical quality of leading models from both countries. However, in terms of enterprise-level model applications or models for less common languages, Chinese companies currently lack the resources to perform specialized optimizations.