US economy: jobs and AI
The US economy expanded only at an annualised 1.4% in Q4 2025, well below forecasts of 3%, the advance estimate of real GDP showed. And over the whole 2025, the US economy expanded 2.2%, or below the 2.4% rate in 2024. The boasting of Trump and his advisers of a US boom doesn’t look quite so convincing after these data. Investment in equipment and software products rose more than 8% in 2025, probably the result of the AI investment boom. Without that, US real GDP growth would have been well under 2% during last year.
US economic growth in 2025 was certainly faster than the other top G7 economies, which are still stagnating with growth rates barely above 1% a year. But an annual growth rate of 2.2% in 2025 is still less half that of China’s and one-third of India’s – although the latter are ‘emerging economies’, not ‘mature’ capitalist economies.
The official consumer price inflation year-on-year rate has been falling during 2025, but several studies have shown that the official measure of consumer price inflation in the US is biased downwards. One study pointed out that mortgage and financing rates are left out of the costs for consumers in the official index. Including such costs would double the inflation rate. In a previous post, I have shown studies that reveal how the official index does not account for the real rise in prices for American households. If it did, using a ‘value rate of inflation’, (that Mino Carchedi and I have developed), it would show that real household incomes have fallen some 20% since 1960 and 4% since the end of the pandemic.
The Trump administration claims that the tariff hikes on imports into the US since last April (now judged illegal by the US supreme court) have had no effect on inflation. But the latest analysis suggests that this is not so. The Congressional Budget Office (CBO) estimates that US companies and consumers are now bearing 95% of the increased tariffs, particularly on imported foodstuffs. The tariffs on food imports have risen 44%, in effect, a 12% tax on consumers. This analysis was backed up by the New York Federal Reserve, which concluded that “US import prices for goods subject to the average tariff increased by 11 percent … more than those for goods not subject to tariffs” and “U.S. firms and consumers continue to bear the bulk of the economic burden of the high tariffs imposed in 2025.” Higher tariffs, labour and health-insurance costs had pushed many businesses to raise prices. “After holding the line on prices for several months, companies—big and small—have begun a new round of increases, in some cases by high-single-digits.”
Headline consumer price inflation on a three-month basis is still 3%. The latest GDP data for Q4 measures what is called the personal consumption expenditure (PCE) index for inflation. This is the measure that the Federal Reserve follows to adjust its policy interest rate. And the core PCE (excluding food and energy) is also above 3% on a 3-month basis (Furman source).
In 2025, retail prices rose faster than average consumer incomes. So stagflation is still in place.
No wonder most American households, apart from the very rich, feel depressed about their lot. It’s what is called a K-shaped economy.
But what about jobs? Does the current relatively low unemployment rate lend the lie to stagnation? The official rate may creeping up, but it is still way below where it was at the end of the pandemic slump.
January’s jobs growth figure looked good on the surface, up 130k. But rises in November and December were revised down. More important, there were even larger revisions to 2025 as a whole, which showed that last year was the weakest year of job creation outside of recessions in two decades. The US economy only added 181,000 jobs in 2025, or just 15,000 extra jobs per month.
Outside of leisure & hospitality, private healthcare, and government – the US economy has been consistently losing jobs.
Manufacturing employment notably fell. So Trump’s economy is not exactly delivering the “manly jobs” that he promised.
Ironically, the health sector desperately needs staff, but Trump’s draconian deportation policy has removed the very immigrant staff that the sector relies on. Immigrant labour has not taken jobs from native-born workers, but instead filled the gaps where native-born labour was unavailable. Foreign-born workers are heavily concentrated in low-paid occupations, like health and social care.
So far, the unemployment rate has not risen too much. But 2026 is likely to change that. Job cut announcements have hit their highest levels since 2009, while vacancies have fallen to 2020 lows. Last month, US employers announced over 108,000 job cuts, the highest start-of-year total since the 2009 Great Recession, with layoffs rising 118% year over year and over 200% from the end of 2025. Many of those cuts were in white-collar professions. Amazon has implemented multiple rounds of cuts, eliminating around 16,000 corporate jobs in January as part of a broader goal to trim some 30,000 white-collar roles. Meta continued layoffs in its Reality Labs division and other teams, with hundreds of positions already cut in early 2026. So the vacancy-to-unemployed ratio is now near 1.0. This is a critical point that will indicate rising unemployment from hereon.
And this brings us to the impact of AI on the economy and jobs. The debate on the impact of AI continues. First, there are the optimists. Stanford University economist, Eric Brynjolfsson predicts that AI will follow a ‘J-curve’ in which initially there is a slow, even negative, effect on productivity as companies invest heavily in the technology, before they finally reap the rewards. And then the boom comes. Such a J-curve was seen in US manufacturing productivity growth, which fell in the mid-1980s and then after the recession of 1991, accelerated sharply until the mid-2000s.
Brynjolfsson now reckons that US productivity rose 2.7% last year, a near doubling from the sluggish 1.4 per cent annual average that characterised the past decade. BJ raised his ‘J-curve’ theory again. “General-purpose technologies, from the steam engine to the computer, do not deliver immediate gains. Instead, they require a period of massive, often unmeasured investment in intangible capital — reorganising business processes, retraining the workforce and developing new business models. During this phase, measured productivity is suppressed as resources are diverted to investments. The updated 2025 US data suggests we are now transitioning out of this investment phase into a harvest phase where those earlier efforts begin to manifest as measurable output.”
But the latest Q4 2025 data throw that prediction into doubt. Another mainstream economist Jason Furman forecasts just a 1.7% rise in productivity for 2025, which would be better than the historic annual average of 1.4%, but a full point below Brynjolfsson. And remember, changes in the productivity of labour depend on two factors: output growth and employment growth. With employment growth near zero in 2025 and 2025 output growth now measured at 2.2%, the maximum rise in productivity will be no more than 2%, and will mainly be due to jobs disappearing, not output rising.
Moreover, is even this current productivity rise in 2025 due to AI? A recent study of how AI adoption affects productivity and employment across more than 12,000 European firms found that AI adoption “increases labour productivity levels by 4% on average in the EU, with no evidence of reduced employment in the short run”. But the productivity benefits, however, were unevenly distributed, benefiting mainly medium and large firms that possess the resources, technical expertise, and economies of scale needed to absorb integration costs. “Moreover, the Ai impact tapered to a one-off effect – rather boosting than long-run productivity growth.”
Another study of 6000 CFOs, CEOs and executives across the US, UK, Germany and Australia found that while 70% of firms actively use AI and two-thirds of top executives regularly use AI, their average use is only 1.5 hours a week, with one-quarter reporting no AI use. And these firms report “little impact of AI over the last 3 years, with over 80% of firms reporting no impact on either employment or productivity”. The executives reckoned that by the end of this decade, AI will boost productivity by 1.4%, increase output by 0.8% and cut employment by 0.7%. This is not anywhere near as large an impact as the optimists expect, and moreover suggests that job losses are just as contributory as any output rises.
According to a new report by the OECD, AI use by companies is rising across OECD countries. In 2025, 20.2% of firms reported using this tool, more than double in comparison to 2023. However, adoption is not the same as productivity. The study shows that AI is like any other technology introduced by capital – it will only be adopted widely if it enables capitalists to shed labour and raise profitability. As Karl Marx observed in Capital, machinery and automation are deployed not to lighten work but to deepen capital’s grip on labour. “The constant aim of these improvements is to diminish the manual labour for a given capital, which, thanks to these improvements, not only requires fewer laborers, but also constantly substitutes the less skilled for the more skilled”. So when the optimists claim a boost in productivity from AI, they are ignoring the impact on jobs and labour’s share of new value created.
Indeed, PIMCO, the giant bond investment managers, reckon if there is any increase in productivity from AI, it is likely to come from shedding jobs rather than from increasing output. In the technology boom of the 1990s, PIMCO says, output rose and workers (at least in the expanding tech sectors) saw real wages rise.
But this time is different. Workers’ real wage growth in 2025 slowed to a pace barely above 1%. This helped drive down labour’s share of U.S. income to its lowest level in decades of reported history.
Any productivity gains from AI will come from job losses and any new value will swing to the owners of AI capital, not workers. Indeed, PIMCO reckons that “AI-related job displacement is already happening, although so far it’s been relatively limited. Entry-level hiring stalled in 2025 in the most AI-exposed sectors. And since 2022, we estimate that cumulative U.S. employment in the most AI-exposed sectors has already declined by over 1% versus a 4% increase in employment across other sectors.”
And more and more studies praising AI do so from the point of view of reducing human labour time. We are told that AI-powered “agents” are cutting the time taken to produce research and forecasting reports. Goldman Sachs recently announced it was working with Anthropicon, an AI agent, to automate roles at the bank. The company says Uber, Netflix, Salesforce and Allianz also use its models in a similar way. PIMCO estimates that approximately one-third of tasks performed by workers across the U.S. economy could be handled by the current LLMs. “If AI hypothetically were to displace only a small 2% share of these workers, it could lead to nearly one million fewer U.S. jobs and the unemployment rate rising by 0.5 percentage points (assuming none of these workers leave the labour force).”
Meanwhile, in 2025, technology investment reached levels unseen in decades. The big 7 tech companies—Google, Amazon, Meta, Microsoft, Nvidia, Apple and Broadcom (8 if one counts Tesla)—have announced $700 billion new spending in AI for 2026 alone, after having invested $450 billion in 2025. Google said it plans to double its capital expenditure this year to as much as $185bn. However, cash flow from existing businesses is increasingly insufficient to meet investment demands from the ‘hyperscalers’ in an industry that requires constant reinvestment, and whose returns may not justify the scale of expenditure. The likelihood of a financial bust remains high. And the jury is still out on the productivity impact of AI.
The US AI tech giants are increasingly banking on achieving super-intelligent models that would transform capitalism – they are searching for this ‘holy grail’ (which remember, was a fiction). The aim is the removal of human labour altogether. Anthropic’s CEO Dario Amodei said that AI could wipe out half of all entry-level white-collar jobs in one to five years. Just this month, Microsoft’s AI chief Mustafa Suleyman predicted most white-collar work “will be fully automated by an AI within the next 12 to 18 months.” The US economy is still “one big bet on AI”.
No comments:
Post a Comment