Artificial Intelligence is Widening the Gap Between Capital and Labour - And the Layoffs Prove It

Meta cut down 8,000 jobs the same week it raised AI spending to $145 billion. It’s not alone - and the pattern reveals exactly where AI’s gains are actually going.

Source: Wikimedia Commons

On April 30, 2026, Meta reported Q1 earnings and raised its full-year capital expenditure guidance to $125-145 billion, largely to fund new data centres and AI infrastructure. Investors, anticipating greater restraint, sent the stock down sharply the next day - a rare pushback against Meta’s spending pace. Three weeks later, the other half of the story arrived: Meta began cutting roughly 8,000 jobs, about 10% of its total workforce, while transitioning 7,000 other employees into AI-focused roles. CEO Mark Zurckerberg told staff cuts were essential because “success isn’t a given” in AI. Meta is not an outlier. It is simply a clear example of a pattern now visible across the entire technology sector: businesses are simultaneously cutting the workforce and raising AI-focused spending, often in the same earnings call. This pattern is the story. AI is not just automating tasks, rather it’s actively redirecting revenue away from workers and towards the infrastructure and shareholders that own the technology, and policymakers have a dwindling window to correct that redirection before it becomes the norm. 

The scale is no longer negligible. Outplacement firm Challenger, Gray & Christmas found that employers tied roughly 101,700 U.S. job cuts to AI in just the first half of 2026 - virtually double the number of all 2025 combined. Oracle cut roughly 21,000 - 30,000 roles. Amazon cut 16,000. Intuit cut 3,000, around 17% of its total staff, explicitly to “reallocate resources towards AI.” IBM, Block, and Cisco each cut thousands more. Challenger, Gray & Christmas also identifies AI as the most-cited reason during the heaviest month of cuts. At the same time, CNBC reports that the four largest hyperscalers - Microsoft, Amazon, Alphabet and Meta - are on track to invest close to $700 billion on AI infrastructure, around 77% higher than 2025’s reported record of $ 410 billion. That is capital, not labour, absorbing the returns AI is supposed to generate: money moving into data centres and narratives that keep investors satisfied, rather than the wages of the workers being cut.

Fig 1. Graph showing hyperscaler capex growth (actual & projected). Source: Bessemer Venture Partners

That said, reports also indicate this may not be the whole picture, and it shouldn’t be treated as one. Part of this may be cases of “AI-washing” - using automation as a convenient explanation for cuts that in reality may be about different issues, such as correcting pandemic-era overhiring or investor pressure to further improve profit margins. OpenAI CEO Sam Altman acknowledged that some companies blame AI for layoffs they would have made anyway. Whether AI is the direct cause or the cover story, the effect remains identical: companies are cutting employee count and increasing capital spending in the same breath, and the workers absorbing the cuts have no claim on the AI revenue being generated. 

This imbalance is the actual inequality risk, and it is a policy failure, not one of law or technology. A corporation moving billions into AI infrastructure while laying off thousands of workers is a corporation with the resources to fund transition support - it’s simply choosing not to, because nothing is currently requiring it to. Three concrete moves could change this reality. First, a windfall tax on AI-driven productivity gains at companies combining record capital expenditure with layoffs, with the proceeds funding retraining that workers can carry between employers, rather than just a one-time severance package. Second, disclosure rules forcing companies to report how much of a layoff is genuine AI-substitution versus cost-cutting, closing the loophole that lets “AI-washing” go unchecked. Third, antitrust regulators should examine what it means for competition when just four companies spend a combined $700 billion on AI in one year. Few firms can approach that scale, concentrating employer power in a handful of companies and eroding the leverage workers would otherwise hold in a more competitive market. 

None of this happens automatically. The 2026 layoff wave shows what happens when it doesn’t: gains begin to concentrate at the top of a balance sheet while the cost lands on workers with the least ability to absorb it. Technology isn’t the deciding factor here. The policy choices that are being made, or not made, while capital expenditure figures climb are.