AI’s Economic Lie

AI is a new economic revolution! It will separate economic growth from labour, enable unlimited growth, and lead to a new age of prosperity. This narrative has been used to justify a truly gargantuan investment in the technology. Last year, Big Tech spent $400 billion on AI. That is enough to eradicate world hunger for a decade! But this year, they are poised to spend $700 billion on AI. Indeed, economists have said these investments have boosted the US economy. Deutsche Bank found that almost all of US GDP growth was from AI investment, and without it, the country would be in a functional recession. Harvard economics professor Jason Furman has backed this up, saying that AI drive 92% of US GDP growth in the first half of 2025. So, is AI the economic miracle it promised to be? Well, no. Economists are starting to actually dig a little deeper past the AI propaganda and have found that this biblical investment has driven 'basically zero' economic growth.

This comes from Goldman Sachs Chief Economist Jan Hatzius. In an interview with the Atlantic Council, Hatzius said that AI spending had 'basically zero' contribution to the US GDP growth in 2025. He even said "We don't actually view AI investment as strongly growth positive" and that "there's a lot of misreporting, actually, of the impact AI investment had on U.S. GDP growth in 2025, and it's much smaller than is often perceived."

So, why does Hatzius have such a different opinion from many of his peers?

Well, unlike them, he actually looks at where all that cash is going. You see, all these billions of dollars are mostly being used to import computer chips. So, Hatzius argues that these investments are adding to the Korean and Taiwanese GDP, not the US GDP. So, in fact, these investments are draining the US of capital it really should be using to grow its own economy!

But Hatzius had a, quite frankly, odd omission. You see, AI is meant to increase productivity through augmentation and automation, right? And, increasing productivity creates GDP growth. So, surely, if AI investment itself isn't increasing GDP, the AIs created by this investment, and deployed into the US economy, will be driving economic growth. So, why didn't Hatzius take this into account?

Well, because AI isn't increasing productivity, and so its deployment isn't increasing GDP.

Analysts have known this for quite a while.

Goldman Sachs itself estimates that AI will increase US productivity by only 15%. But even that might be wildly optimistic. Back in 2024, ING estimated that AI would only increase productivity by 1%. For some context, the computer and internet revolution accounted for roughly 45% of the US's productivity gains since 1980, and from 1980 to 2019, US productivity increased by roughly 60%, meaning PCs and the web drove a 27% productivity increase. Even during the peak IT craze in 2000, the industry's annual capex was $164.2 billion, or $310 billion in today's money. So, even optimistically, AI is costing more than twice what the IT revolution did, and will only deliver half the productivity gains. More realistically, it is set to deliver less than 4% the productivity gains.

But I think ING's predictions are actually too optimistic, as recent data makes it clear AI is not at all the productivity tool we once thought it was.

A study from Carnegie Mellon University found that even the best "agentic" AIs completely fail basic tasks 70% of the time. A recent study found that the best current AIs failed 97.5% of realistic real-world freelancing jobs given to them. METR found that generative AI slowed down experienced developers by 19%, as they have to spend more time correcting mistakes the AI had made than the time the AI saved them. All of this explains why a recent Harvard Business Review report found that AI is not boosting productivity, but instead intensifying work. In short, AI is more like a burnout machine than a productivity tool.

We can see these findings reflected at the business level too. Famously, MIT found that 95% of corporate AI pilots fail. This is backed up by PwC, who found that only 12% of businesses using AI saw it reduce costs and increase profit, BCG, who found that only 5% of companies that deployed AI saw value from it, and Forrester Research, who found that 15% of their corporate survey correspondents reported an increase in profit margins from AI over the past year.

As such, it looks like businesses are pulling back from AI use, rather than rapid adoption. The Economist found that the use of AI by large US corporations is actually shrinking, not growing, and S&P Global Market Intelligence found that the cancellation rate of corporate AI programs skyrocketed from 17% in 2024 to 42% in 2025.

Again, we can see all of this reflected in economic data. Despite what mainstream media might have told you, there are currently zero layoffs caused by AI automation or augmentation, according to Oxford Economics.

So, we shouldn't be surprised that Hatzius didn't even bother to discuss how AI deployment could increase productivity and, therefore, GDP, because the data clearly shows that simply isn't happening.

But, it is actually worse than that, because despite the colossal amount of cash being poured into AI, it isn't going to get much better than it already is.

OpenAI's recent research paper found that increasing the computing power behind AI, or shoving more data into them, can't reduce AI "hallucinations" or make AI more accurate than it is today. In fact, they found there is no viable way to reduce AI hallucinations, heavily implying these models are doomed to stay as unreliable as they currently are. This is corroborated by a paper written by the eminent Vishal Sikka and his son Varin Sikka, which claims to mathematically prove that these AIs "are incapable of carrying out computational and agentic tasks beyond a certain complexity." All of this also backs up research into the efficient compute frontier, the scaling problem, and the Floridi conjecture, which detail these limitations of AI.

So, just because big tech is shoving $700 billion into AI over the next year, it doesn't mean it is going to get any better at all. That is why I think ING's 1% productivity estimate is actually unrealistically optimistic.

But this means that the very justification for that insane $700 billion investment is totally wrong. It isn't going to make AI properly useful, it isn't going to boost productivity, it isn't going to drive GDP growth, it isn't the economic silver bullet it was promised to be.

So, what does that say about where this is all going? This is a dead end. Shoving that cash into it is just going to make this economic bubble even bigger and make the fallout when it pops even worse.

The more I look at the AI industry, the more it is starting to look like a boys' club. They are just passing money between themselves to make each other richer, and giving blatantly crap and hollow excuses to the rest of us to explain it away. There are other words for this, like circular financing or an oligarchical economy. But one thing is for sure, it isn't going to end well.

Thanks for reading! Everything expressed in this article is my opinion, and should not be taken as financial advice or accusations. Don't forget to follow me here or support me over at Substack.

(Originally published on PlanetEarthAndBeyond.co)