TechFlow News, March 18: According to Business Insider, NVIDIA CEO Jensen Huang proposed during his keynote address at GTC 2026 that AI tokens—the fundamental billing unit for large language model text processing—should be incorporated into engineers’ compensation structures. He stated that Silicon Valley has already begun treating “how many tokens are allocated per role” as a key factor in hiring negotiations and believes engineers with token allocations will see significantly higher productivity.
OpenAI CEO Sam Altman envisions a broader framework. During the All-In podcast in May 2024, he suggested the future may bring a “Universal Basic Compute” model, under which every citizen receives an allocation of GPT-7 compute capacity—usable personally, resellable, or donatable for purposes such as cancer research. This month, Altman also remarked at BlackRock’s Infrastructure Summit that AI model providers’ core business will shift toward “selling tokens,” akin to public utilities like water and electricity.
Microsoft CEO Satya Nadella cautioned that if mass-produced tokens fail to tangibly improve efficiency in healthcare, education, and the public sector, society will struggle to sustain tolerance for the massive energy consumption of AI data centers.
