Are AI tokens the new signing bonus or just a cost of doing business?

This week, a topic that has been boomeranging around Silicon Valley was in the spotlight. It is an AI token as a reward. The idea is very simple. Instead of just providing engineers with salaries, assets, and bonuses, companies also provide them with a budget of AI tokens, the computational units that power tools like Claude, ChatGPT, and Gemini. Use it to run agents, automate tasks, and explore your code. The bottom line is that access to more compute makes engineers more productive, and highly productive engineers become more valuable. The idea is that it is an investment in the person who holds it.

Nvidia’s leather-jacketed CEO Jensen Huang seemed to capture everyone’s imagination at the company’s annual GTC event earlier this week when he floated the notion that engineers should get about half their base salary back in tokens. By his calculations, his top management could be burning through $250,000 a year in AI computing costs. He called it a recruiting tool and predicted it would become standard across Silicon Valley.

It’s not entirely clear where the idea first came from. Tomasz Tunguz, a prominent Bay Area VC who runs Theory Ventures, focuses on AI, data, and SaaS startups, and has gained a loyal following over the years through his writing about all things data, spoke about this in mid-February, writing that tech startups are already adding inference costs as a “fourth component of engineering compensation.” Using data from compensation tracking site Levels.fyi, he puts the top quartile software engineer salary at $375,000. Adding $100,000 of tokens will fully load you with $475,000. That means roughly one in five dollars is now spent on computing.

That’s no coincidence. Agentic AI gained popularity and the conversation accelerated significantly with the launch of OpenClaw in late January. OpenClaw is an open source AI assistant designed to run continuously while the user sleeps, switching tasks, spawning subagents, and processing to-do lists. This is part of a broader shift toward “agent” AI. This means a system that not only responds to prompts, but also performs a series of actions automatically over time.

The practical result is that token consumption has exploded. If a person writing an essay can spend 10,000 tokens in an afternoon, an engineer running a swarm of agents can spend millions of tokens a day automatically, in the background, without you having to type a word.

By the end of this week, the New York Times had compiled a smart look at the so-called tokenmaxxing trend, finding that engineers at companies including Meta and OpenAI were competing on internal leaderboards that track token spending. Generous token budgets are quietly becoming a standard job perk, much like dental insurance or free lunches once were, the newspaper reported. An Ericsson engineer in Stockholm told the Times that although his employer would take responsibility, he would spend more on Claude than he would earn in salary.

Perhaps tokens will actually become the fourth pillar of engineering rewards. But engineers may need to draw a line before accepting this as a simple victory. More tokens may mean more power in the short term, but it doesn’t necessarily mean more job security given how quickly things are evolving. First of all, large token allocations come with high expectations. If your company is effectively funding the computing value of a second engineer on your behalf, the implicit pressure is to produce at twice (or more) the rate.

And underneath that, there’s a more confusing problem. At the point where a company’s token spending per employee approaches or exceeds that employee’s salary, the financial logic behind headcount starts to look different from the finance team. When computing is doing the work, it becomes harder to avoid the question of how many people need to coordinate.

East Coast-based Jamaal Glenn, a Stanford MBA and former VC turned financial services CFO, similarly points out that what appears to be a benefit can be a clever way for a company to inflate the apparent value of a compensation package without raising cash or assets. This actually benefits employees over time. Your token budget is not set in stone. No thanks. It won’t show up in your next offer negotiation in the same way as base salary or stock grants. Once companies successfully normalize tokens into salaries, it may be easier to maintain cash compensation while pointing to increasing computing allowances as evidence of investment in employees.

It’s a good deal for the company. Whether it’s a good deal for engineers depends on questions that most engineers don’t yet have enough information to answer.