Firms Shift Payroll to AI Tokens, Returns Lag

Companies moved salary budgets into per-token AI compute and agent software while reducing headcount in 2026, but surveys and company reports indicate the changes often did not deliver better financial returns or customer service.

The four largest cloud providers guided roughly $700 billion combined in capital expenditure for 2026. Analyst estimates put AI agent software spending near $207 billion this year, a 139% increase from the prior period. Data tracking U.S. layoffs shows AI as the most-cited reason for job cuts for a fourth consecutive month, and technology made up 31% of first-half layoffs.

At the close of a major industry conference, Nvidia’s chief executive proposed a test that compares an engineer’s salary to their annual token use, saying a $500,000 engineer who consumes less than $250,000 in tokens would be a concern. Nvidia is planning toward a $2 billion yearly token bill for its engineering teams. Several large employers shifted budget lines previously used for salaries to pay for per-token compute and agent subscriptions.

A Gartner survey of 350 executives at companies with more than $1 billion in revenue that are deploying AI agents or automation found roughly 80% had cut headcount, but the survey did not find a clear link between those headcount cuts and improved financial returns. Gartner analyst Helen Poitevin commented, “Workforce reductions may create budget room, but they do not create return.”

Company examples show limits to the token strategy. A ride-hailing firm supplied 5,000 engineers with AI coding tools and used its full 2026 AI budget by April. About 70% of committed code was AI-generated, and company leadership reported the work had not yet produced measurable improvements in customer-facing products. The company capped engineers’ AI spend at $1,500 a month after the budget overrun. Other large employers introduced similar per-user token caps when demand exceeded forecasts.

A European fintech replaced roughly 700 customer service roles with an AI assistant, paused human hiring for more than a year and emphasized the approach to investors. Customer satisfaction fell and complaints rose after the change. The fintech’s chief executive acknowledged the company had focused heavily on efficiency and cost and said the outcome was lower quality; the company has since rehired human support staff.

Gartner expects that by 2027 half of companies that cut customer service staff for AI will rehire, often under different job titles. In a separate poll of 321 customer service leaders, Gartner found only 20% had genuinely reduced staffing because of AI, suggesting many reductions reflected cost control rather than full automation.

A 2026 AI Index from a research center at Stanford University found employment for software developers aged 22 to 25 fell nearly 20% from 2024, while older developer cohorts continued to grow. Analysts note the token-versus-salary calculation varies by market: the example that compares token budgets to a $500,000 engineer applies to a small share of U.S. engineers and is less relevant in regions where local labor costs are lower and token charges can be higher.

Finance teams responded to early token overruns by capping usage rather than restoring layoffs. Several chief financial officers imposed monthly spending limits after AI budgets were exhausted. Companies reporting improved return on investment tended to be those that invested in staff who use AI tools instead of fully replacing people with automation.