Mercor’s chief executive says the startup now spends more on AI tokens than on employee salaries, underscoring how quickly artificial intelligence costs are reshaping the economics of some businesses. Brendan Foody, Mercor’s CEO, said he would bet that “in five years the average enterprise spends more on compute than headcount,” according to Business Insider.
The remark reflects a broader shift in the AI industry, where companies are increasingly funneling money into the computing power needed to run and train models rather than into traditional labor costs. That trend is already visible across the sector: Uber has reportedly capped employee AI spending after blowing through its budget in just four months, while other companies are expanding their AI-related investments and revising forecasts upward as demand accelerates.
Mercor’s spending pattern is notable because it suggests AI is not just a tool for boosting productivity, but also a major operating expense. For startups and large enterprises alike, the cost of tokens — the units of text AI systems process and generate — can rise quickly as usage scales, especially when staff rely heavily on AI tools for coding, writing, analysis, or customer support.
The comment also fits into a larger debate over whether AI will ultimately replace or simply augment workers. Bloomberg has reported on companies across industries that are using AI to cut costs, raise output, and redesign workflows, including private-sector efforts inspired by government cost-cutting campaigns. At the same time, some firms are discovering that widespread AI adoption can create new budget pressures rather than immediate savings.
For businesses watching Mercor’s model, the question is whether heavy spending on compute will produce enough efficiency gains to justify it. That issue is becoming central as more companies race to build AI products, secure data-center capacity, and manage rising infrastructure bills, even as investors continue to reward firms that show clear revenue gains from AI.
What happens next will depend on whether Mercor and similar companies can turn those token costs into durable growth. Foody’s forecast suggests the balance between labor and compute may keep shifting, with AI infrastructure becoming a larger line item in corporate budgets and a more important factor in hiring, pricing, and profitability.
