Baidu CEO Emphasizes Ongoing Need for Data Center and Cloud Infrastructure Investment Despite AI Cost Efficiency Challenge
DUBAI, United Arab Emirates - Despite the skepticism raised by DeepSeek's cost-effective large AI models, Baidu CEO Robin Li maintains the necessity for investment in data centers and cloud infrastructure.
"Cloud infrastructure investment is still crucial," Li stated at the World Government Summit. "To create models superior to others, we require more compute."
Compute encompasses the hardware resources that drive AI models, enabling data training, information processing, and prediction generation.
DeepSeek's breakthrough has garnered global recognition for its language models that rival leading systems like GPT with significantly lower computing consumption. This has led to questions about the necessity of substantial AI infrastructure spending.
Baidu was an early adopter of AI products after OpenAI's ChatGPT release. However, its Ernie model, which Li claims matches GPT-4, has had limited public traction.
Li acknowledges the unpredictable nature of innovation, citing DeepSeek's emergence. He also recognizes the stimulative impact of U.S. chip sanctions on Chinese innovation within computing constraints.
Additionally, Li has modified his stance on closed-source development, now recognizing the potential of open-source approaches to accelerate AI adoption. "By opening things up, we spark curiosity and facilitate faster technological dissemination."