Baidu CEO Robin Li stated on Tuesday that investment in data centres and cloud infrastructure remains crucial despite the challenge posed by Chinese AI startup DeepSeek. Speaking at the World Government Summit in Dubai, Li emphasised that smarter AI models require increased computing power, or "compute," to function effectively. His comments come as DeepSeek has gained attention for creating language models that perform similarly to OpenAI's GPT while using much less computing power, prompting debate over the need for large-scale AI infrastructure.

Baidu, a key player in China's AI development, was quick to launch its own AI products after the release of OpenAI's ChatGPT in late 2022. However, its own language model, Ernie, has seen limited public adoption, despite claims that it rivals GPT-4 in capability. Li, who previously argued that no OpenAI-like company would emerge from China, admitted at the summit that innovation in AI is unpredictable, as shown by DeepSeek's rapid rise.

In a shift from his earlier stance on AI development, Li acknowledged that open-source models could play a significant role in accelerating AI adoption. While he had previously advocated for closed-source approaches, he now recognises that allowing greater access could foster wider experimentation and faster technological spread. This marks a notable change in Baidu's approach to the evolving AI landscape.