Goldman Sachs is telling displaced tech workers what the labor market has been whispering for months: the jobs lost to artificial intelligence are not coming back on the same terms. Workers pushed out by automation face longer searches and smaller paychecks when they do land, the bank’s research warns, even as headline employment figures still look passable.

The real story is not the layoff count. It is the quality of re-employment. When high-skill workers absorb permanent earnings damage, it signals a structural shift in the labor market that feeds directly into consumer spending, credit quality, and the kind of slow-burn demand destruction that eventually shows up in safe-haven flows and metals pricing.

Goldman strategist Pierfrancesco Mei laid out the math in stark terms. Displaced tech workers take roughly one month longer to find new work than peers in more stable fields. And when they do get hired, they earn less. As the Daily Mail reported, Mei wrote that these workers “suffer real earnings losses of more than 3 percent upon re-employment, compared with negligible losses for workers displaced from more stable occupations.” Three percent may sound modest. Over a career, compounded across hundreds of thousands of displaced workers, it is not.

The numbers behind the warning

In March alone, U.S. employers announced 60,620 job cuts, a 25 percent increase from the previous month. Artificial intelligence was linked to roughly one in four of those cuts. That ratio matters. It means AI is no longer a background force in labor-market data. It is becoming a named cause.

Related Posts
Retirement crisis looms with 46% of Americans at zero

The company-level detail tells the same story from a different angle. Oracle cut over 30,000 jobs globally, with more than 10,000 positions axed in a single morning on March 31. Amazon has slashed tens of thousands of corporate roles across multiple rounds while doubling down on AI spending. Meta launched another wave of layoffs this year. Atlassian cut around 10 percent of its workforce to refocus on AI. Block, Jack Dorsey’s payments company, axed 40 percent of its workforce, with more than 4,000 employees let go in a single day. Epic Games trimmed 1,000 jobs amid rising costs and slowing growth.

These are not marginal trims at failing companies. They are structural reorganizations at profitable, cash-rich firms that are choosing to redirect capital from human labor toward machine intelligence. The distinction matters for anyone trying to read the macro cycle.

The headline jobs number hides the damage

The Bureau of Labor Statistics reported that total employment rose by 178,000 in March and that the unemployment rate held largely unchanged at 4.3 percent. On the surface, that looks healthy enough. But the surface is doing a lot of work.

In February, American employers unexpectedly cut 92,000 jobs. March’s rebound to 178,000 new positions papered over that hole, but the composition of those jobs and the quality of re-employment are where the real signal lives. Goldman’s research suggests that workers who lose positions to automation are not simply sliding into equivalent roles. They are being pushed into lower-level or more routine work, which erodes their earning power and, over time, their household balance sheets.

This is the kind of labor-market deterioration that does not flash red on a dashboard. It shows up gradually, in weaker consumer spending, softer credit metrics, and rising delinquencies among formerly middle-class borrowers who quietly took a pay cut they did not choose. As Jamie Dimon’s recent annual letter warned, systemic financial risks often build in exactly these kinds of slow, structural channels before they become visible in headline data.

Why metals investors should pay attention

Gold does not move in a straight line from labor-market stress to higher prices. The transmission mechanism is indirect but real. When high-earning workers absorb permanent income damage, the downstream effects ripple through consumer demand, housing, and credit quality. If enough of these displacements accumulate without offsetting job creation at comparable wages, the result is demand destruction that monetary policy eventually has to address.

The Federal Reserve watches labor-market conditions closely when calibrating interest-rate policy. A labor market that looks adequate on the unemployment rate but is quietly hollowing out in quality gives the Fed room to cut rates or hold them lower for longer. Lower real rates, all else equal, reduce the opportunity cost of holding gold. That is the basic plumbing.

But there is a second channel. When workers displaced by technology cannot find equivalent roles, and when major banks are publicly warning about the problem, it erodes confidence in the durability of the expansion. That kind of confidence erosion feeds safe-haven demand. It also raises questions about the sustainability of the equity valuations built on the very AI spending that is driving these layoffs. As we noted in our recent look at S&P 500 dividend yields, big tech is already paying almost nothing back to shareholders while commanding record valuations. If the AI buildout produces productivity gains but also produces a weaker consumer, the math gets harder to defend.

Structural displacement vs. cyclical layoffs

The distinction Goldman is drawing deserves close attention. Cyclical layoffs happen when demand softens and companies trim headcount to protect margins. Those workers typically get rehired when the cycle turns, often at comparable wages. Structural displacement is different. It occurs when the job itself disappears or when the skills that commanded a premium become less valuable. The worker may find new employment, but the terms are worse.

Mei’s finding that displaced tech workers suffer real earnings losses of more than 3 percent while workers from more stable occupations suffer negligible losses points squarely at structural displacement. Automation did not just reduce headcount. It reduced the market value of certain human skills. That is a harder problem for policymakers to solve, and it carries longer-lasting economic consequences.

For metals investors, structural labor-market shifts matter because they tend to be disinflationary on the demand side while potentially inflationary on the policy-response side. If displaced workers spend less, demand weakens. If policymakers respond with fiscal transfers, deficit spending, or easier monetary conditions, the currency takes the hit. Either path can be constructive for gold, though through different mechanisms and on different timelines.

The current environment, where major banks are simultaneously warning about pullback risk in gold and about structural labor damage, reflects a market caught between competing forces. The macro case for hard assets keeps strengthening even as short-term positioning gets crowded.

The AI arms race as a macro variable

What makes this cycle unusual is the speed and scale of the corporate pivot. Amazon, Oracle, Meta, and their peers are not cutting jobs because revenue collapsed. They are cutting jobs because they believe AI can do the work cheaper and faster. The capital is not leaving the system. It is being redirected from labor to machines.

That redirection creates winners and losers. The winners, for now, are the companies selling AI infrastructure and the shareholders of firms that successfully cut costs. The losers are the displaced workers and, eventually, the consumer economy that depends on their spending. The question for markets is whether the productivity gains from AI arrive fast enough to offset the demand drag from displaced workers earning less.

Goldman’s research suggests the transition is not smooth. The one-month longer job search and the 3-plus percent earnings haircut are friction costs that aggregate into real economic drag. And the layoff pace appears to be accelerating, not stabilizing. March’s 25 percent jump in announced cuts over the prior month, with AI cited as a factor in one of every four, points to a trend that has not peaked.

In a world where geopolitical stress is already elevated and energy-supply disruptions remain a live risk, adding structural labor-market deterioration to the mix creates a more fragile macro backdrop. Fragility, over time, is what drives capital toward assets that do not depend on someone else’s balance sheet or someone else’s employment contract.

What to watch next

Several things will determine whether Goldman’s warning stays a research footnote or becomes a defining feature of this cycle:

  • Re-employment quality: If displaced tech workers continue absorbing earnings losses, the demand drag compounds. Watch consumer-credit data and spending trends in tech-heavy metro areas.
  • Layoff pace: March’s 60,620 announced cuts and the AI attribution rate of roughly 25 percent are the benchmarks. If those numbers climb in coming months, the structural thesis strengthens.
  • Fed response: A labor market that looks adequate on the unemployment rate but is deteriorating in quality gives the Fed a complicated signal. Easier policy in response would likely support gold prices.
  • Equity valuations: If AI spending drives layoffs but does not deliver proportional revenue growth, the valuation case for tech weakens, potentially redirecting capital toward hard assets and amplifying the kind of volatility that has already rattled gold markets this year.

The quiet erosion

Goldman’s warning is not about a crisis. It is about something harder to trade and harder to ignore: a slow, structural downgrade in the earning power of a large class of skilled workers. The unemployment rate will not capture it. The payroll headline will not flag it. But household balance sheets will feel it, and eventually, so will the credit cycle.

For investors focused on capital preservation, the lesson is familiar. The most dangerous risks are the ones that do not announce themselves. They accumulate in the gap between what the headline number says and what the underlying economy is actually doing. Gold has always been, at its core, a bet that the gap will eventually matter.