The tech industry is rapidly considering a new component of engineer compensation: direct access to AI compute. Companies are beginning to allocate budgets for AI tokens – the units used to power models like ChatGPT, Claude, and Gemini – alongside traditional salary, equity, and bonuses. The rationale is simple: greater access to compute increases engineer productivity, making them more valuable. This is not just a fringe idea; Nvidia CEO Jensen Huang recently suggested engineers should receive roughly half their salary again in tokens, potentially reaching $250,000 annually for top performers.
The Rise of “Tokenmaxxing”
This shift is driven by the explosive growth of “agentic” AI, where systems don’t just respond to prompts but autonomously execute tasks over time. Tools like OpenClaw, an open-source AI assistant, exemplify this trend: running continuously, spawning sub-agents, and processing tasks without constant human input. As a result, token consumption is soaring. Engineers running AI agents can burn through millions of tokens daily, a stark contrast to the 10,000 tokens someone might use for a single writing task.
The New York Times recently reported that engineers at Meta and OpenAI are now competing on internal leaderboards tracking token usage. Generous budgets are becoming standard, mirroring perks like dental insurance or free meals. One Ericsson engineer in Stockholm reportedly spends more on AI compute than their entire salary, with the company footing the bill.
Why This Matters
This trend reveals a fundamental change in how tech companies measure productivity. Compute is increasingly being treated as a direct input into an engineer’s value, rather than an indirect cost of operation. This is significant because it shifts the focus from hours worked to outputs generated, potentially rewarding high-intensity AI-assisted work over other contributions.
However, there are downsides. The implied expectation of doubled productivity with increased token allocation creates pressure. More critically, as token spend approaches or surpasses an engineer’s salary, companies may begin to reassess headcount. If the AI is doing the work, the need for human coordination becomes a financial question.
A Question of Value
Financial experts like Jamaal Glenn, a former VC and CFO, point out that tokens are not a replacement for cash or equity. Token budgets don’t vest, appreciate, or carry weight in future negotiations. Companies may use this system to inflate compensation packages without increasing actual long-term employee value. The move could allow them to keep cash compensation flat while marketing a growing compute allowance as investment in their workforce.
Ultimately, whether this new model benefits engineers remains to be seen. The current lack of transparency and long-term implications raise critical questions that employees must address before fully embracing AI tokens as a legitimate part of their pay.
