Good news! The massive investment into AI has made American households on average significantly wealthier, and will continue to do so for years (if not the next decade). Bad news? The gains will only reinforce the K-shaped economy in the medium term; any improvement in the fortunes of medium to low earners is some way off.
AI has delivered more than a 7% uplift in household wealth for U.S. consumers, Oxford Economics CEO Innes McFee told the company’s Global Economic Outlook conference in London this week. However, this “powerful boost” has mostly landed in the pockets of high-income Americans.
The “wealth effect” created by the blockbuster spending in AI (households feeling wealthier because the value of their assets is increasing, and thus increasing their spending) will reinforce the K-shaped economy, likely until 2035, McFee later told Fortune in an exclusive interview. The K-shaped economy is a phenomenon in which the fortunes of the wealthy track steadily higher, while those on the lower end of the income spectrum gradually sink.
In 2025 the idea of this diverging economy was supported by research from the likes of Moody’s chief economist Mark Zandi, who observed that the economy is being almost exclusively driven by the sentiments of the “well-to-do.”
And while AI does have the potential to one day help close the inequality gap, living standards will continue to fork before then.
Asked if the AI boom will reinforce the K-shaped economy for decades to come, McFee answered: “Absolutely. Eventually, AI may well end up being a driving force to bring those two groups a bit closer together, but in order to see that, you need to see the productivity gains at a low level. The productivity gains at a low level, low-skilled jobs have to come through because that means … real wages increase, and that’s ultimately what drives living standards.”
He added that unification may not happen in the next five or even 10 years: “Eventually, it might bring things together, but in the meantime, through the wealth effect, through investment and all those sorts of things, it’s unlikely that AI helps at all with the K-shaped economy.”
Echoes of the K-shaped economy can be traced back over decades: The Fed began monitoring the distribution of household wealth in Q3 2010, and reported that total wealth equaled $60.76 trillion. Of that, the top 0.1% owned $6.53 trillion, and the top 99% to 99.9% owned $10.75 trillion. By contrast, the bottom 50% shared only $330 billion.
Fast-forward to Q3 2025: The wealth of the bottom 50% has grown by 1,189% to $4.25 trillion—though still significantly behind the wealth held by the top 0.1% even some 15 years prior. The top 0.1% saw their wealth grow 281% to $24.89 trillion, nearly six times the wealth held by the bottom 50% combined.
‘Hollowing out’ the middle-skill jobs
Modeling suggests that AI adoption across businesses is likely to be something of an S-curve, beginning slowly and then rapidly increasing before gradually leveling off. Per Oxford Economics’ modeling, there will never be full integration of AI in businesses, as it can’t be used to replace physical trade jobs.
As a report from the Penn Wharton Budget Model observed last year, AI adoption will plateau in the early 2030s, the result of declining opportunities to employ additional AI tools productively.
The protection from AI for trade jobs (and indeed, the potential boon data centers represent for the likes of plumbers and electricians) does mean a “hollowing out” of certain roles will occur in the coming years, McFee said.
“We saw that a little bit after the financial crisis for different reasons,” McFee tells Fortune. “You’ll see lots of employment growth at the lower end of the distribution and right at the top, but in the middle, maybe even a contraction in job growth. That’s largely because those middle-skilled jobs are the ones where you can largely substitute tasks away with something like AI. You need critical analysis, you need the ability to question things—that tends to come at the top end of distribution. The middle tends to be filled with people who are just learning those skills.”
