The most persistent fear in the AI conversation, that robots are coming for everyone’s jobs, just took a hit from actual corporate spending data. Firms that invested heavily in AI tools grew their headcounts by 10.2% over two years, according to a new working paper. The ones that barely spent anything on AI? Their employment numbers basically flatlined.

The study, published June 30, 2026 by Ramp Economics Lab, analyzed spending records from 21,559 US firms and cross-referenced them with workforce data from Revelio Labs. Co-authored by Ara Kharazian and Ryan Stevens of Ramp alongside Lisa Simon from Revelio Labs, it offers one of the most granular looks yet at what happens to payrolls when companies actually open their wallets for AI.

The spending gap tells the story

The study divided firms into two buckets based on their AI spending intensity. High-intensity adopters spent an average of $33.67 per employee on AI tools. Low-intensity adopters spent $2.78 per employee.

The high spenders saw that 10.2% headcount increase over two years. The low spenders showed no statistically significant change in employment levels.

Perhaps the most counterintuitive finding involves entry-level positions. Instead of declining, entry-level headcount at high-intensity firms grew by 12%, outpacing even the overall hiring bump.

Why this study hits different

What makes this paper stand out is the data source. Rather than relying on surveys, job postings, or executive sentiment, the researchers used actual corporate spending records from Ramp’s platform.

The findings also align with previous academic work, though they paint a more aggressive picture. Research from MIT Sloan had previously suggested an approximate 6% increase in employment growth linked to AI adoption, but that was measured over a five-year period. The Ramp study found a 10.2% increase in just two years.

What this means for investors

The competitive landscape is also shifting. The data suggests a widening gap between firms that commit to AI integration and those that treat it as optional, with high-intensity spenders significantly outpacing low-intensity spenders in hiring dynamics.

The risk to watch is whether this correlation holds as AI tools mature. The current data covers a period of early adoption, when companies are likely layering AI on top of existing workflows rather than restructuring around it.