Nvidia’s chief executive has a blunt message for anyone predicting mass AI unemployment. “The second idea is AI is going to destroy jobs. That’s also completely nonsense,” Jensen Huang told the Dreamforce 2026 keynote in San Francisco on September 15, appearing alongside Salesforce chair and CEO Marc Benioff. His argument: AI lifts productivity, lets companies attempt projects that were previously impractical, and the resulting ambition creates demand for more work rather than fewer jobs.
It is the kind of claim that used to be safe to file under vendor optimism. What makes this round of it different is that the 2026 hiring data is — patchily, imperfectly, but measurably — moving in the direction Huang and Benioff describe, even while the layoff data keeps a darker count on the same ledger.
What the keynote actually claimed
Benioff’s version was softer than Huang’s. AI is more likely to generate new work than eliminate jobs, he said, describing companies using AI to change how employees interact with business applications, access corporate data, and offload routine tasks. Salesforce used the keynote to launch AI Force, a platform for connecting AI models with enterprise data — the corporate frame for the claim being that AI deployed inside a company’s own systems augments workers rather than replacing them. Both positions are arguments from companies that sell AI, a caveat worth holding onto throughout.
What the layoff data says
The most-cited counterpoint comes from Challenger, Gray & Christmas, whose monthly job-cut report has been tracking AI as a named cause of layoffs since 2023. The 2026 tally through August: AI has been cited in 116,175 US job cut announcements, about 22 percent of all cuts, and it remains the leading cited reason year-to-date.
But the monthly trend bends the other way. August’s AI-attributed cuts fell to 3,462 — the lowest monthly total since December 2025 and the first month since February in which AI did not lead the reasons, edged out by restructuring. August’s overall total of 52,881 cuts was the quietest August since 2022.
Challenger’s own commentary also cuts against reading the AI number too literally. “Naming AI in a layoff announcement can win over investors while pushing current and prospective employees away. That’s why the messaging has swung from hedging to aggressively citing it,” Andy Challenger said in the July report. The firm has flagged that attribution is genuinely murky: a Bronx hospital system that cut utilization-review nursing positions after adopting new software was tracked as “technological update (possibly AI)” because the hospital never confirmed the product was AI. Both the headline AI number and Huang’s dismissal of it sit on data that neither side of the debate fully controls.
What the hiring data says
Here the keynote’s optimism has more footing. Challenger’s tally of announced hiring plans reached 107,500 workers through July, up 25 percent from the same period in 2025 and the strongest January-July total since 2023. Technology — the industry “gorging on AI talent,” in LinkedIn economists’ words — leads all sectors in announced 2026 hires at just over 19,000.
ZipRecruiter’s 2026 AI Employer Report, published in July, put survey numbers on the same shift: 35 percent of employers say AI will increase their total headcount going forward, and 24 percent say that has already begun. PwC’s 2026 AI Jobs Barometer, analysing over a billion job ads, found headcount growth of 52 percent at the most AI-exposed companies versus 36 percent at the least exposed, with wage growth of 24 percent against 17 percent. Individual AI roles show the same pull: LinkedIn’s platform data has AI engineer postings up 156 percent between 2024 and 2025, and Gen Z workers make up more than two-thirds of hires into forward deployed engineer and AI engineer roles.
That is not the whole labour market. The same ZipRecruiter report found 38 percent of employers have shifted basic data processing away from entry-level workers and onto AI, and 31 percent have raised experience requirements for entry-level jobs as a result. Glassdoor chief economist Daniel Zhao’s assessment of the tech sector specifically is blunt: tech employment has dipped 4 percent from its 2022 peak, and he has suggested the sector’s job opportunities are worse than in the 2008 crisis. The New Zealand data centre build-out offers a local illustration of the same dynamic — enormous construction-phase employment around AI infrastructure, with far fewer permanent roles once facilities open, as our earlier analysis of the BCG jobs numbers covered.
And the CEOs preaching job creation have their own records on the other side of the ledger. Salesforce announced around 4,000 job cuts in 2025 while restructuring around AI, with Benioff saying at the time that AI-layoff framing was being used as a scapegoat. Nvidia, for its part, grew headcount through the same period. Neither fact settles the argument; both are worth knowing before taking either CEO’s word for it.
Reading the two ledgers together
The honest synthesis is that both numbers are true at once. Companies are restructuring around AI and naming it in roughly 116,000 cut announcements this year — while simultaneously planning more hires than at any point since 2023 and paying double the median salary to the people who build with the technology. Challenger’s August data even shows the AI-attribution wave cooling just as the investor-messaging incentive to name AI peaks — a reminder that the layoff ledger measures announcements, not necessarily mechanisms.
For workers, the practical signal in the combined data is less about whether Huang is right in the aggregate and more about where the demand is concrete: AI-adjacent implementation roles are growing fast enough that the forward deployed engineer has become the breakout job title of 2026, and the wage premium for AI skills keeps widening rather than narrowing. The displacement side of the ledger is equally concrete for clerical, data-processing and entry-level work — a split ZipRecruiter’s data captures and the keynote rhetoric, from either side, tends to flatten.
FAQ
What did Jensen Huang say about AI and jobs? At Dreamforce 2026 in San Francisco on September 15, Nvidia’s CEO said predictions that AI will destroy jobs are “completely nonsense,” arguing AI raises productivity and creates demand for new work. He made the remarks alongside Salesforce CEO Marc Benioff.
How many jobs cuts has AI been blamed for in 2026? Challenger, Gray & Christmas has recorded 116,175 US job cut announcements citing AI from January through August 2026 — about 22 percent of all cuts, and the leading cited reason year-to-date.
Are AI-exposed companies hiring or cutting? Both. PwC’s 2026 AI Jobs Barometer found 52 percent headcount growth at the most AI-exposed companies versus 36 percent at the least exposed, while Challenger counts AI as the top named reason for job cuts. The effects are concentrated in different roles: growth in AI-skilled positions, cuts in clerical and data-processing work.
Is AI increasing or decreasing total employment? The data supports both directions depending on segment: 35 percent of employers in ZipRecruiter’s survey expect AI to raise headcount, while 38 percent have moved basic data processing off entry-level staff. Tech sector employment overall remains about 4 percent below its 2022 peak.
— CJ Murden, editor of Singularity.Kiwi. Former digital technologies teacher, author of AI-focused books. Writing with a New Zealand focus.
