OpenAI CEO Sam Altman told a Commonwealth Bank of Australia conference in Sydney on May 26 that rapid AI development would not trigger a global “jobs apocalypse,” saying he was “delighted to be wrong” about how quickly AI would displace white-collar workers. The statement landed the same week Meta finished notifying 8,000 employees of layoffs tied to its AI pivot, and less than a month after Oracle cut up to 30,000 workers while redirecting capital toward AI data centers.

What Altman Said

“I don’t think we’re going to have the kind of jobs apocalypse that some of the companies in our space advocate or talk about,” Altman told CBA Chief Executive Matt Comyn, according to Benzinga. He said OpenAI leaders had been “roughly right” about the technology’s trajectory since ChatGPT’s 2022 debut but “pretty wrong” about social and economic impacts.

Altman described testing AI-generated replies for Slack and email, including messages labeled “this is Sam’s AI,” before deciding to personally handle sensitive conversations again. “We really do care about our interactions with people,” he said, adding that the experience changed his view of how AI could reshape employment.

The Layoff Receipts

The timing creates a sharp contrast. On May 20, NPR reported that Meta notified 8,000 employees of layoffs while shifting 7,000 others into AI-focused teams. TIME reported that Oracle cut up to 30,000 workers in the past month as the company redirects billions toward AI infrastructure, with some employees saying they were asked to document workflows to train AI systems before being let go. Oracle will be cash flow negative until at least 2030, according to Bloomberg data cited by TIME.

A May 18 analysis by ReadUncut examined SEC filings for 33 companies that publicly linked layoffs to AI between 2023 and Q1 2026. The total: roughly 155,000 positions eliminated. The finding: for the vast majority of those companies, operating margins either stayed flat or declined after the cuts. Meta’s revenue surged 33% but its operating margin slipped from 41.5% to 40.6%, “because the infrastructure required to run AI at Meta’s scale consumes the savings faster than the savings accumulate,” the analysis found. The only companies with improving margins were those selling AI infrastructure to everyone else.

The IPO Context

Altman’s reassurance comes as OpenAI reportedly filed IPO paperwork confidentially in May 2026, targeting a public listing in September or October with discussions around a potential $1 trillion valuation. Companies approaching IPO have strong incentive to frame their technology as productivity-enhancing rather than job-destroying, because regulatory and public sentiment around workforce displacement can complicate the offering narrative.

The Productivity Question

The gap between C-suite rhetoric and financial outcomes is now documented in audited filings. Altman may be right that the “apocalypse” framing is overblown: displacement has been slower, more targeted, and less catastrophic than 2023-era predictions suggested. But the companies invoking AI to justify layoffs are, by their own SEC disclosures, not seeing the margin improvements they promised. The jobs are gone. The returns, for most, are not there yet.