Nvidia CEO Jensen Huang (젠슨 황) has pointed to “tokens” as a core resource in the artificial intelligence era, stressing that the benchmark for technological competition is changing.

On March 22, the South China Morning Post reported that Nvidia recently declared at its GTC event that it is shifting beyond being a simple chip supplier to becoming an “AI factory designer.” It explained that the standard product produced by this factory is tokens. Here, tokens are not cryptocurrency but the smallest unit of data that AI models process or generate, and a computing resource.

Huang likened tokens to “oil in the AI era,” and predicted that token output, pricing and efficiency will determine corporate and national competitiveness. Global big tech companies are investing huge sums to expand data centres and computing infrastructure to secure token production capacity.

Against this backdrop, China’s strategy is also drawing attention. China is expanding AI competition from a chip-centred approach to a token economy based on “power plus computing.” Alibaba is shifting into an AI-focused company, reorganising its internal operations around a token hub, and aiming to expand market share by leveraging the price competitiveness of domestic models. Many of the top models by global token consumption are made in China, showing strength in cost efficiency.

China’s strength, in particular, is energy. As the world’s largest power producer, China is pushing computing-power synergies as a national strategy, based on expanding renewable energy such as solar and wind. This structure combines data centres with energy supply to lower token production costs.

As the axis of AI competition moves from chip performance to token production capacity, the likelihood is growing that the United States and China will compete for leadership through different approaches: technology-driven for the United States and power-based for China.