100 AI agents per worker: A productivity windfall for some, not all

Jensen Huang’s vision of a future with 100 AI agents per worker signals a major shift from simple generative AI to fully agentic systems that can plan and execute tasks. While this model could unlock significant productivity gains, it depends heavily on infrastructure—data centers, electricity, semiconductors, and supply chains. Global constraints in power, materials, and geopolitics mean adoption will be uneven. Advanced economies and well-capitalized firms will benefit first, while others may lag. The result is not just a technological transition, but a structural divide shaped by access to compute, energy, and industrial capacity.