Digital Human Industry Enters Shakeout Phase After Large Model Boom

In 2021, various digital humans began appearing in live streams, advertisements, and short video dramas. The meteoric rise of "Liu Yexi" briefly made digital humans a hot trend. However, after the emergence of large models, the number of companies discussing this concept has decreased, replaced by talk of AI agents and AI assistants.

An entrepreneur in the AI virtual companionship field told Jiemian News that the core reason for the declining number of digital human companies in the market is their lack of AI capabilities. Especially in the 3D digital human sector, the large model boom has actually raised the bar for companies.

A report released by the Zhongshang Industrial Research Institute shows that in 2024, 2D digital humans were implemented relatively quickly, capturing a 70.1% market share. 3D digital humans held a smaller share of 29.9%. The core reason is that 3D digital humans remain constrained by the pace of technological iteration.

A CEO of an LED display manufacturer that has collaborated with multiple digital human companies told Jiemian News that a phenomenon exists in the digital human industry: while top-tier companies have excellent products, the overall industry product quality is uneven. Furthermore, high GPU costs and unstable services make it difficult to achieve error-free 7x24 hour service.

However, digital human companies face bottlenecks in both data and talent. For instance, regarding data, text-to-image and text-to-video data are now relatively abundant, but 3D data involving human motion and environmental interaction remains scarce, and most companies lack the capability to accumulate such data. Additionally, Chai Jinxiang mentioned that talent specializing in 3D content and AI research has traditionally been "siloed," making it challenging to bring these two types of professionals together.

Platformization is contingent on market acceptance. An industry insider told Jiemian News that currently, many companies offering digital human solutions are focusing their implementation in areas like exhibition halls, cultural tourism sites, and education and training scenarios, where the tool-oriented attribute is stronger. This is closely tied to policy support—since the start of this year, the release of strategies encouraging AI development by various regions has prompted many institutions to prioritize digital applications.