Chinese tech giants turn to home-grown AI to transform everyday digital services
Chinese consumer technology companies are quietly developing AI models to improve delivery, travel, gaming, video and social platforms.
China’s artificial intelligence industry is attracting global attention through companies such as DeepSeek, Z.ai, Moonshot AI and MiniMax. However, another group of technology companies is developing AI systems with a less visible but increasingly practical focus: integrating artificial intelligence into the everyday services used by millions of consumers.
Table Of Content
Rather than operating as dedicated AI laboratories, these companies span e-commerce, social media, video streaming, gaming, food delivery and travel. They are developing foundation models designed around their own platforms, business data and customer needs. The approach reflects a wider shift in China’s technology sector as companies seek to make AI a core part of their existing operations rather than simply offering it as a standalone product.
Consumer platforms are building their own AI systems
RedNote, the lifestyle and social media platform often compared with Instagram, recently highlighted the growing trend with the launch of Dots3-Note Preview. Developed by the company’s AI research arm, Dots Studio, the open-weight model has 280 billion parameters and is designed to handle a broad range of tasks.
Dots Studio said benchmark testing showed the model could match or outperform systems from US companies including OpenAI and Anthropic, as well as Chinese AI developers such as DeepSeek and Z.ai, in certain tasks. The release represents a notable expansion of RedNote’s technology ambitions beyond its established role as a platform for fashion, travel, shopping and lifestyle recommendations.
In a blog post, Dots Studio said it aimed to “build AI that benefits everyone and helps people solve the many problems they encounter in life”. The development illustrates how consumer platforms are increasingly treating AI as a core part of their technology rather than an additional feature supplied by outside providers.
Su Lian Jye, chief analyst for applied intelligence market research across the Asia-Oceania region at Omdia, said the investment reflected a broader change in how non-AI companies approach artificial intelligence. “When non-AI companies invest in AI models, it is primarily because AI has become critical to their business operations,” Su said.
The companies involved also have an advantage because they operate large digital platforms that generate substantial amounts of proprietary information. Su described the consumer platforms as “digital natives and extremely data-rich”. Combining that business data with customised AI models can help companies improve internal workflows, anticipate demand, personalise services and develop additional sources of revenue.
AI is spreading across travel, delivery, gaming and video
The same strategy is appearing across several major areas of China’s technology industry. Meituan, one of the country’s largest food delivery and local services platforms, has invested heavily in its LongCat model family. The company is seeking to apply AI across a wide range of operations, including merchant services, logistics, customer support and recommendations.
Travel is another area where companies are developing specialised AI technology. Trip.com, China’s largest online travel agency, has created its own travel-focused AI system, Wendao. The system is designed to support services such as itinerary planning, customer assistance and personalised travel recommendations.
These applications demonstrate why companies may prefer to develop models that understand their particular industries. A general-purpose AI system can perform a wide range of tasks. Still, a model trained or adapted for a specific business can potentially make better use of the company’s own information and processes.
Video and gaming companies are also experimenting with similar approaches. Chinese video platform Bilibili has developed the Index family of language models, including IndexTTS, a multilingual speech synthesis system. The technology has been applied to tasks such as AI-generated subtitles, translation, video production and voice generation.
Gaming developer miHoYo has taken a different approach with Glossa, an in-house generative AI model intended to make game characters more interactive. The technology has reportedly been used for AI features in Honkai: Star Rail, the company’s internationally successful role-playing game.
miHoYo is also examining how artificial intelligence could support the wider game-development process. Potential applications include generating content, producing synthetic voices and improving production workflows. For companies producing large volumes of digital content, these uses could reduce repetitive work and let development teams focus on more complex creative tasks.
High costs and data challenges remain
Despite the potential benefits, developing an in-house AI system is not straightforward. Companies must invest in computing infrastructure, skilled employees and data while also ensuring that AI systems comply with increasingly complex regulatory requirements.
Omdia’s AI Market Maturity Survey identified several major obstacles for businesses adopting artificial intelligence. These include insufficient or poor-quality business data, regulatory and compliance requirements, shortages of qualified in-house staff and difficulties connecting AI systems with existing technology.
Data can be particularly important for consumer platforms. Companies may have access to large amounts of information, but that does not necessarily mean the data is suitable for training or operating AI systems. Data may need to be cleaned, organised and governed before it can provide useful results, adding further technical and financial demands.
Infrastructure costs are another concern. Training and operating large AI models requires significant computing resources, while specialised AI teams can be expensive to recruit and retain. Su warned that investment in artificial intelligence could “introduce budget pressure due to high infrastructure and personnel costs”, although “the long-term impact can be strongly positive”.
The growing number of companies developing proprietary AI systems suggests that artificial intelligence is becoming less of a standalone technology race and more of an operational tool. While headline-grabbing AI laboratories continue to compete over increasingly powerful models, China’s large consumer platforms are quietly concentrating on how the technology can improve the services people already use every day.
