For many months, corporate America painted Artificial Intelligence as the next industrial revolution. The tech giants are pouring in 100s of billions into data centers, research, and chips. President Donald Trump recently hailed this AI investment as the reason why the US economy is the hottest worldwide. However, beneath these soaring stock prices and headlines, economists and industry analysts are questioning if the AI boom is delivering anything real to the American economy. This disconnect between companies spending and country’s gaining, as per them, has never been wider. But is this all true? Or are they worried just for nothing?
America receives a truth bomb about its GDP
As Jan Hatzius, Goldman Sachs chief economist, recently sat down with the Atlantic Council, he stated clearly about the state of economic impact of AI. As per reports, the United States has recorded 2.2% GDP growth in 2025. However, Hatzius states, Artificial Intelligence contributed “basically zero” to the number. He said, “We don’t actually view AI investment as strongly growth positive.” With it, he directly challenged the narrative that drove tech stocks to the record highs.
Hatzius further suggested that much of what politicians and the press have reported about the economic contribution of AI is simply wrong. He explained, “I think 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.” The problem is not that the companies are not spending—they are doing that absolutely. It is where that money ends up, actually.
An analyst on Wall Street put the spending frenzy in a stark perspective on X. As per the analyst, “OpenAI is forecasted to burn $218 billion in cash from 2026 to 2029.” For comparison, $18.2 billion was burned by Uber over 6 years, before turning profitable. Netflix in 7 years and Tesla in 11 years burned $11.1 billion and $9.3 billion, respectively. The analyst noted that, “OpenAI plans to burn more than all three combined in less than four years.”
This comparison clearly reveals something quite uncomfortable about the AI investment’s current wave. Tesla, Uber, and Netflix were building up things that had clear paths to revenue—cars people can buy, rides that could be booked, and even subscriptions that people will pay for. By contrast, OpenAI is burning cash at an unprecedented rate while they are, as per reports, losing money on $200 monthly subscription.
If not America, who is cashing the big checks?
The press releases do not tell people that the AI boom in America is not real, as most of the money does not stay here. When Amazon, Microsoft, Meta, or Google announce massive data center expansions and infrastructure plans, the media just celebrates the tech leadership of America. However, the GDP map tells a completely different story. As per it, AI is ruining things and not improving.
Joseph Briggs of Goldman Sachs explains that about three-quarters of Big Tech’s projected capital expenditure flows to Asia’s manufacturing economies directly. Those fancy chips that are powering the AI dreams of America are largely made in South Korea and Taiwan. When any US company buys semiconductors worth a million dollars from any Taiwanese firm, that particular transaction adds to the GDP of Taiwan and not America.
Hatzius even told Atlantic Council that, “The equipment powering AI is imported.” “A lot of the AI investment that we’re seeing in the U.S. adds to Taiwanese GDP, and it adds to Korean GDP but not really that much to U.S. GDP,” he further added. Such a reality check cuts against all the things like soulless AI slop tech executives and politicians have been selling the public.
What truly makes this awkward for the cheerleaders of the industry is that even spending that stays domestic is not translating to any productivity gains. The recent survey of approximately 6000 executives across the US, Australia, and Europe found that while 70% firms are using AI actively, 80% reported having absolutely no impact on productivity or employment. The companies are forcing their employees to use the tech, but the results are not showing up in a measurable way.
Slow technical progress brings in more rising skepticism
The economic questions are quite bad, but the technical challenges facing AI development might even be much more troubling. The entire investment thesis of the industry is resting on the assumption that throwing more computing power at the problems will yield dramatic improvements. However, this assumption is now starting to crack.
The AI scientists are now acknowledging that the “more compute equals better results” approach is not working like it previously did. A jump from GPT-3 to GPT-4 was huge. This leap justified massive investments and bold predictions. However, a jump ahead was much smaller. Even improvements are showing some diminishing returns, with a huge increase in computing power.
For the investors who are pouring billions into the space, it creates an uncomfortable paradox. The tech that is supposed to justify all spending is now improving slowly. The costs, on the other hand, associated with those improvements are now exploding. An X user captured a growing skepticism with a biting summary of it is definitely AI that failed so far.
As per the user, “In exchange for getting workers fired, stealing artists’ work, harming the environment, giving people psychosis, hallucinations, and creating CSAM, you will get, uh, well…” This post clearly points to a broad unease that extends beyond economics. It makes us question what exactly society is truly getting in return for all this spending and disruption.
