At an AI-focused investment summit in Beijing, Lenovo Group Vice President and General Manager of China Infrastructure Business Group, Chen Zhenkuan, shared his outlook on the rapid evolution of AI computing infrastructure. Speaking during a panel discussion centered on the theme of identifying certainty in the AI infrastructure era, he highlighted the escalating demands placed on foundational computing power as large language models continue to expand in scale.

Chen pointed to the trajectory of model growth, noting that OpenAI's ChatGPT initially launched four years ago with 175 billion parameters, while DeepSeek's model released in early 2025 boasts 671 billion parameters. More recently, Kimi unveiled a system in July of this year that pushes parameter counts toward the trillion-level mark, with enterprises continuing to plan even larger models on the horizon.

He explained that supernodes, which are typically defined as systems with 32 or more GPU cards, are becoming essential. As model parameters grow, traditional 8-card GPU servers are increasingly unable to independently handle the computational load of larger-scale models, making it necessary to rely on supernodes offering 32-card or higher configurations for adequate computational support. "This is not a matter of choice, but a mandatory component of AI infrastructure," Chen said.

Looking at the industry landscape, Chen noted that roughly five AI chip companies currently support supernode capabilities, and he anticipates that leading AI chip makers will broadly embrace the technology by next year. "Next year will certainly be a defining year for supernodes," he said. "If this year has been about demonstrating feasibility, next year we will see large-scale commercial deployment."

Chen added that as supernodes transition into commercial application, the momentum will ripple across the broader industry chain, spurring development, reshaping the competitive landscape, and giving rise to new standout enterprises within the sector.

These remarks were delivered during a roundtable discussion at the AI Investment Summit, where industry leaders convened to explore the most reliable opportunities emerging in the AI infrastructure space.