May 29, 2026 – The 2026 World Intelligence Industry Expo kicked off on Wednesday at the National Convention and Exhibition Center in Tianjin, and one message from the stage stood out above the rest: every company will soon need its own AI factory.
Jiang Tao, NVIDIA’s vice president of global operations, took the spotlight during the expo’s AI-themed day to outline what he calls the dawn of the “Token Economy” — a shift he argues will redefine how businesses think about computing, infrastructure, and value creation.
“We are no longer building data centers to store and compute,” Jiang said. “We are building factories that produce intelligence.”
The transformation, according to Jiang, rests on three pillars.
First, the compute foundation is flipping. For decades, CPUs ruled IT architecture. Now, GPUs — once the domain of graphics rendering — have become the backbone of AI infrastructure. Enterprise workloads are migrating en masse from traditional computing to accelerated computing, and the gap is widening fast.
Second, AI itself is evolving beyond anything the industry has seen. What started as perceptual AI has given way to generative models, and the next frontier — agentic and physical AI — demands exponentially more from inference complexity, model scale, and real-time responsiveness. AI is no longer a research experiment, Jiang emphasized. It is becoming a production-grade productivity system that must be woven into actual business operations and industrial workflows.
Third, and perhaps most strikingly, the output of these new facilities is changing. Legacy data centers delivered storage, compute power, and bandwidth. The AI factories of tomorrow will output tokens, intelligence, and tangible business outcomes. Jiang drew a direct analogy: just as traditional factories take raw materials and electricity to produce cars or steel, AI factories consume power and data to produce intelligence, decision-making capability, and efficiency.
The implication is sweeping. Jiang stated plainly that demand for AI infrastructure is no longer confined to tech giants or internet-native companies. Manufacturing, finance, energy, healthcare, research, and government — every sector is entering a full-scale AI transformation.
“In the next decade, AI will be the new fundamental productive force,” Jiang said, drawing a parallel to how past industrial revolutions brought electricity, machinery, and the internet. “AI factories will be as common as data centers are today. Tokens will be as essential as electricity. And intelligence will be the most critical resource any enterprise can hold.”
