Move from content to execution
With rapid development in AI large models and agent technology, the digital human market is set for broader commercial application. Third Bridge experts expect digital human use cases to expand gradually from today's focus on content display and interaction in live streaming and short video, into execution-oriented settings such as enterprise services, education, customer service and intelligent assistants. The technology path has changed markedly. During the rise of the metaverse concept, virtual humans built on 3D modelling held the mainstream market. As generative AI and large model technology developed rapidly, digital humans based on 2D video cloning gradually became the market mainstream. Third Bridge experts note that 2D digital humans are currently more widely used in live streaming, short video and customer service, because they come closer to a real person's presentation and can already support real-time interaction. By comparison, 3D digital humans are used more in games, film and television, and large-scale virtual brand endorsement.
On market applications, experts say the areas with heavier digital human use include bank customer service, telecoms operator service systems, offline showrooms and short video content production. Of these, live streaming and short video remain the largest market.
Enterprise applications may become an important direction for further growth. Experts believe digital humans will gradually enter corporate workflows, used for staff training, new joiner onboarding, human resources management and corporate content safety training.
The virtual endorser market is also drawing attention. Experts say virtual humans offer advantages over real celebrity endorsement in lower cost, controllable risk and long-term usability, and may in future attract more companies to build their own digital brand identities.
Experts note, however, that digital human technology still faces several challenges. Among them, the ability to match movement to spoken content remains insufficient. Digital human live streaming can currently synchronise voice and lip movement, but a gap remains between physical performance and context. Experts believe this is mainly constrained by the maturity of video generation models, compute costs and inference capability. Future progress will require training on more high-quality data alongside improved real-time generation capability.
On the future competitive landscape, experts believe what really determines a company's competitiveness is price, compute efficiency and practical application capability. The core purpose of a digital human is to replace repetitive labour at lower cost. If product pricing approaches the cost of human labour, commercial value will be limited. The key to industry development therefore lies in reducing compute consumption and solving efficiency problems in real commercial settings.
Looking ahead, experts believe the range of digital human applications will widen further, including automotive intelligent assistants, robot interaction and autonomous driving. As AI technology continues to develop, digital humans may shift from being content production tools towards intelligent assistants with the capacity to execute.
All insights in this article are based on information provided by Third Bridge experts.
For media enquiries, please contact: comms@thirdbridge.com
Transcript references:
1. 2026/06/16 Silicon Intelligence’s Pre-IPO – Striding Towards the First Listed Digital Human Stock on the HKEX (Conducted in Mandarin)
This analysis is also available in Chinese via Third Bridge's WeChat account.
