Companies Spending the Most on AI Are Hiring the Most People

If AI were eating entry-level jobs, you’d expect the companies buying the most of it to be hiring the fewest people. According to a study cited by economist Anders Humlum in an NPR report published on 18 August 2026, the opposite is happening.

🔍 THE BOTTOM LINE — The firms spending the most on AI tools grew entry-level headcount by 12% in the two years after adoption, per a Ramp and Revelio Labs analysis of more than 21,000 US firms. That doesn’t mean AI anxiety is baseless — but it does mean the loudest narrative in the graduate jobs debate may have the causation backwards.

The Study Behind the Claim

The research, conducted by financial technology firm Ramp with workforce analytics company Revelio Labs, examined AI spending and employee headcount across more than 21,000 US firms from early 2021 to early 2026. Its finding, as Humlum summarised it: at companies making the largest AI investments, entry-level head count grew by 12% over the two years following AI adoption.

“If we look at the heavy users of these tools, the firms that are paying a lot of money to Anthropic and OpenAI to subscribe to their models, they are hiring more than anyone else,” Humlum told NPR.

Humlum is a University of Chicago economist who has published some of the most-cited research on AI’s actual workplace effects, so his framing carries weight. His broader point: if AI were straightforwardly replacing entry-level roles, heavy AI spenders should be shrinking. They are doing the reverse.

Why the Grad-Jobs Story Feels Worse Than the Data

None of this means young job seekers are imagining things. The New York Fed puts unemployment for recent US graduates — 22 to 27-year-olds with a new bachelor’s degree or higher — at 5.7% as of June, versus 4.1% for all workers. A ZipRecruiter survey found 47% of recent graduates believe AI has already affected hiring in their field.

But the economists NPR interviewed disagree about how much AI deserves the blame:

Stanford’s Erik Brynjolfsson says AI is “part of the story and the evidence is building.” His payroll research found early-career workers aged 22 to 25 in AI-exposed roles saw a 16% relative employment decline since late 2022. His explanation is structural: language models are trained on codified, written-down knowledge — exactly the kind of knowledge new graduates bring, and exactly what senior workers replace with experience.
Harvard’s David Deming points at timing instead. The decline in junior hiring started roughly six months before ChatGPT launched, which suggests something else is driving it. His candidate: remote work. Companies are less willing to hire and train juniors into remote roles, and a New York Fed analysis found remote job growth tracks graduate unemployment more closely than AI adoption does.
Humlum points at the spending data. Firms that buy AI heavily are growing junior headcount, not cutting it.

All three agree a real shift is underway. Where they part ways is on whether AI is currently the cause of graduate hiring pain — or the most convenient explanation for it.

What This Means for Job Seekers

The honest reading of the evidence is that AI is reshaping what entry-level work looks like faster than it is deleting entry-level jobs. Snowflake’s March 2026 global survey with Omdia — 2,050 business and technology leaders across 10 countries — reported 77% of organisations experiencing AI-driven hiring against 46% reporting role reductions, with IT operations, cybersecurity, and software development seeing the strongest net job creation. That is a vendor-commissioned survey, so treat the exact figures with appropriate caution, but the direction matches the Ramp and Revelio headcount data.

For graduates, the practical signal across all three economists’ work is consistent: the knowledge that gets a degree is the part AI handles; the judgment, client work, and tacit skills built on the job are what employers still pay humans for. Roles that get you into the room — even imperfect ones — matter more than ever, because the alternative is competing with a model on the one thing models do best.

FAQ

Is AI really causing the graduate jobs problem? Economists genuinely disagree. Stanford research shows a real employment decline for young workers in AI-exposed roles, but Harvard’s David Deming and the New York Fed argue remote work explains more of the timing, and the Ramp/Revelio data shows heavy AI spenders hiring more juniors.

Which fields are gaining AI-related jobs? The Snowflake/Omdia survey points to IT operations, cybersecurity, and software development as the strongest net job creators, and PwC’s AI Jobs Barometer reports jobs requiring AI skills growing far faster than the overall jobs market.

What should new graduates do differently? Target roles that build on-the-job judgment rather than codifiable tasks, and prioritise employers that are investing in AI — the evidence suggests they are the ones expanding, not contracting.

NZ Angle

New Zealand graduates face the same structural question from the other side of the hemisphere. SEEK NZ’s employment reporting has shown AI skill mentions in local job ads roughly doubling year on year, sitting ahead of Australia as a share of listings — meaning NZ employers, like their US counterparts, are paying a premium for AI-adjacent capability while entry-level competition intensifies. The Humlum finding offers a useful filter for Kiwi graduates choosing between employers: the organisations investing seriously in AI are, on current evidence, the ones still growing their junior ranks.

Sources