This week, a topic that has been boomeranging around Silicon Valley bounced into the spotlight: AI tokens as compensation.<\/p>\n
The idea is straightforward enough \u2014 rather than giving engineers only salary, equity, and bonuses, companies would also hand them a budget of AI tokens, the computational units that power tools like Claude, ChatGPT, and Gemini. Spend them to run agents, automate tasks, crank through code. The pitch is that access to more compute makes engineers more productive, and that more productive engineers are worth more. It\u2019s an investment in the person holding them, is the idea.<\/p>\n
Jensen Huang, the leather-jacket-wearing CEO of Nvidia, seemed to capture everyone\u2019s imagination when he floated the notion at the company\u2019s annual GTC event earlier this week that engineers should receive roughly half their base salary again \u2014 in tokens. His top people, by his math, might burn through $250,000 a year<\/a> in AI compute. He called it a recruiting tool and predicted it would become standard across Silicon Valley.<\/p>\n It isn\u2019t entirely clear where the idea was first, well, ideated. Tomasz Tunguz, a renowned VC in the Bay Area who runs Theory Ventures and focuses on AI, data, and SaaS startups \u2014 and whose writing on all things data has garnered a loyal following over the years \u2014 was talking about this in mid-February, writing that tech startups were already adding inference costs as a \u201cfourth component<\/a> to engineering compensation.\u201d Using data from the compensation tracking site Levels.fyi, he put a top-quartile software engineer salary at $375,000. Add $100,000 in tokens and you\u2019re at $475,000 fully loaded \u2014 meaning roughly one dollar in five is now compute.<\/p>\n That\u2019s no coincidence. Agentic AI has been taking off, and the release of OpenClaw<\/a> in late January accelerated the conversation considerably. OpenClaw is an open-source AI assistant designed to run continuously \u2014 churning through tasks, spawning sub-agents, and working through a to-do list while its user sleeps. It\u2019s part of a broader shift toward systems that don\u2019t just respond to prompts but take sequences of actions autonomously over time.<\/p>\n The practical consequence is that token consumption has exploded. Where someone writing an essay might use 10,000 tokens in an afternoon, an engineer running a swarm of agents can blow through millions in a day \u2014 automatically, in the background, without typing a word.<\/p>\n