The integration of DeepSeek into Tencent and Baidu’s ecosystems marks a significant shift in China’s AI landscape. By embedding DeepSeek into WeChat and Baidu Search—two of China’s largest digital entry points—the AI model now influences hundreds of millions of users. This development highlights both the competitive and cooperative dynamics between rising AI players and established internet giants.
On one hand, the adoption of DeepSeek reflects the challenges traditional internet giants face in keeping up with advanced AI models.
DeepSeek’s superior performance, cost-efficiency, and usability have outshined many in-house AI efforts, forcing companies like Tencent and Baidu to integrate third-party solutions to stay competitive. On the other hand, these tech giants are strategically leveraging DeepSeek to reinforce their existing ecosystems, transforming external AI capabilities into their own defensive moat.
At its core, this trend underscores the inevitable restructuring of the internet landscape in the AI era. The battle is not merely about AI supremacy but about how established platforms can adapt and maintain dominance by integrating disruptive technologies.
Reshaping Ecosystems in the AI Era
Since the advent of generative AI, a persistent challenge has been the deep contextual integration of AI models into real-world applications. AI’s potential has been constrained by data silos across different platforms, making it difficult to provide truly personalized and contextualized services.
According to an IDC report, 73% of China’s internet industry data is concentrated in the hands of Tencent, Alibaba, ByteDance, and other top platforms.
However, cross-platform data sharing remains an issue, with successful data interconnectivity below 5%. This fragmentation means that user behavior data from e-commerce giants like Taobao and JD.com remains isolated from social media platforms such as Douyin and Weibo.
Similarly, consumer habits on Meituan, video-watching behavior on Douyin, and financial activities on Alipay exist in separate ecosystems, limiting AI’s ability to construct comprehensive user profiles.
Search functions present a similar challenge. While AI models such as Kimi and Zhipu can access online information, most third-party AI models are restricted to public data and cannot access private or proprietary content.
Tencent’s own AI assistant, Yuanbao, is one of the few capable of tapping into exclusive content from its ecosystem, such as WeChat Official Accounts.
Baidu, as China’s largest search engine with over 700 million users, offers a rich multi-functional ecosystem spanning search, maps, forums, cloud storage, and healthcare.
User queries on Baidu often reflect more than simple information retrieval—they represent multi-faceted interactions requiring integrated AI capabilities. The ability to connect these disparate services and leverage the underlying data is an advantage deeply embedded within platforms like Baidu and Tencent, making it difficult for third-party AI models to achieve full integration.
However, despite controlling vast amounts of data, traditional internet companies have struggled to develop AI models that fully meet user needs. Ranking data from AI benchmarking platforms like Sinan shows that Tencent’s Hunyuan and Baidu’s Ernie trail behind DeepSeek-R1 in terms of model performance.
This has led to a paradoxical scenario: while AI firms excel in model capabilities, they lack large-scale user entry points, while tech giants, despite their vast ecosystems, lag behind in AI innovation.
DeepSeek’s Disruptive Influence
The introduction of DeepSeek-R1 has disrupted this status quo. With high performance at a low cost, DeepSeek has quickly expanded its partnerships across industries including smart hardware, automotive, media, internet services, and semiconductors.
This rapid adoption underscores a growing realization among industry leaders: in the AI race, only a few dominant models will likely emerge as market leaders, with DeepSeek being a top contender.
The strategy adopted by Tencent and Baidu—leveraging third-party AI technology while maintaining control over user ecosystems—bears similarities to Google’s approach with Android. While Android remains open-source, Google maintains dominance over the global Android ecosystem through its proprietary services like Google Play, Gmail, Maps, and YouTube.
Similarly, Chinese tech giants may use external AI models like DeepSeek as foundational tools while reinforcing their ecosystems through proprietary services.
Yet, a critical question remains: will this strategy be as effective in the AI era as it was in the mobile era? As AI technology becomes more homogenized, will ecosystem-driven advantages continue to define industry leadership?
The Battle Between Platforms and AI Models
If Tencent and Baidu can maintain their dominance simply by integrating superior third-party AI models, would they eventually abandon their in-house AI research? The answer appears to be no.
Shortly after integrating DeepSeek, Tencent updated its AI assistant Yuanbao and launched a new inference model, Hunyuan-T1, in a limited test phase.
According to Tencent, Hunyuan-T1 is designed for complex reasoning tasks, leveraging WeChat Official Account content to enhance response accuracy and timeliness. This move signals that Tencent remains committed to developing its own AI capabilities despite adopting DeepSeek.
Ultimately, the evolving relationship between AI startups and established internet giants will shape the future of China’s digital economy.
While collaborations like Tencent and Baidu’s adoption of DeepSeek may seem like a temporary concession, they also reflect a longer-term strategy: using external AI advancements to bolster their own ecosystems while continuing to invest in proprietary AI development. The outcome of this competitive yet cooperative battle will determine the dominant players in the AI-driven internet landscape.