Investment in AI has not just outstripped government-led initiatives like the Manhattan Project and the Apollo program. It has also exceeded the capital put into other, more recent, market-driven technology booms.
The dotcom frenzy of the late 1990s and early 2000s saw billions flow into speculative internet startups, and the cryptocurrency boom of the late 2010s and early 2020s attracted its own wave of venture capital. Yet, both pale in comparison to the AI spending wave, which has concentrated greater sums into a shorter timeframe.
This unprecedented spending spree is reshaping global technology and drawing Wall Street’s focus to a handful of giants, but it also creates risks of a financial bubble and questions about the circular financing deals propping up skyrocketing valuations.
What exactly is all this AI money being used for? While some is funding software development, including improvements to the large language models (LLMs) that underpin the technology, much of this capital influx is being used to create physical infrastructure like data centers, energy generators and even computer chips.
According to Stanford University’s AI Index Report, companies spent $37 billion in global private investment on AI infrastructure alone in 2024.
Most investment in AI is going toward infrastructure
Major areas of global private corporate investment in AI
Natural Language
processing
Natural Language
processing
Natural Language
processing
Natural Language
processing
A line chart showing yearly investments in different sectors of AI from 2018 onwards with most investments moving towards AI infrastructure in 2024
Note: Only sectors with over $5 billion invested in any year are shown
Source: Stanford University HAI AI index report 2025
Over 500 new large data centers have been built between 2021 and 2024, and they can be the size of multiple American football fields. Although this construction boom can benefit local economies, concerns have mounted over their massive use of land, water and power.
A McKinsey report from April estimated that a $5.2 trillion investment in data centers will be needed by 2030 to meet the worldwide demand for AI alone.
Billions of AI investment dollars have also gone into healthcare, autonomous vehicles, financial technologies and manufacturing, as a broad range of industries seek to utilize generative AI to increase their productivity.
Both large, pre-existing corporations like Microsoft and Alphabet as well as startups are getting in on the AI boom. In just the first quarter of this year, AI startups raised over $70 billion globally, capturing nearly 60% of all venture capital financing.
This has sent the valuations of AI-focused companies through the roof. Since the release three years ago of ChatGPT, the advanced AI chatbot developed by OpenAI, the value of Nvidia, the company that develops processors that are widely used for running AI applications, has increased more than 10 times hovering near $4.5 trillion. OpenAI, with a $500 billion valuation, is now considered the world’s most valuable private company.
However, big valuations do not necessarily mean big profits. While revenues from AI firms have risen, they have lagged far behind the valuations.
The chart shows the change in market capitalizations and revenues from 2022 to 2025 for three leading AI companies- Nvidia, Microsoft and OpenAI
OpenAI is governed by a nonprofit and operates as a public-benefit company, not a typical corporation
Source: LSEG data, news reports
“That is not an uncommon thing when there is new technology. Stock prices are forward looking, so they don’t necessarily need positive cash flows now to recognize something as valuable,” says Carnegie Mellon University finance professor Bryan Routledge.
Some skeptics have called AI a bubble or at least suggest it’s too soon to say if the technology will truly be worth the investments going into it.
AI supporters have said it’s already transformed our economy for the better and will only increase productivity.
As much as money is pouring into AI, the race for AI development has also become a cash guzzler for the tech companies themselves. The Magnificent Seven tech companies — Google parent Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla — are expected to spend more than $300 billion on AI in 2025.
The cash these tech firms are pouring into AI is making them even more attractive to Wall Street. The cost of building ever-larger AI systems is growing, so stock buyers have increasingly gravitated toward the handful of companies seen as best placed to capitalize on the technology. The Magnificent Seven have outpaced the rest of the S&P 500 by nearly 15 percentage points since January 2025.
However, a concern among investors has been the circular nature of the financing among some tech firms involved in AI investments. Several of these companies are investing in or lending to their own customers so they can continue making purchases from them.
Such deals arouse suspicion among investors since they create an illusion of organic growth with interconnected companies using the same initially invested capital to purchase each other’s products, thereby inflating company valuations.
A recent pullback in tech stocks has underscored the fragility of these financing loops. Oracle bonds slipped after reports it will take on another $38 billion to fund AI infrastructure, adding to more than $100 billion in existing debt as spending outpaces operating income to secure clients like OpenAI.
“The circular deals to me are like a warning sign saying that there isn’t really enough genuine revenue from real customers,” said Andrew Odlyzko, an emeritus University of Minnesota mathematics professor who studies financial bubbles.
Despite these concerns, AI investment is on track to match or exceed the level of investment in some of the biggest technology buildouts in history, such as shipping canals, railroads and telecommunication networks.
Alongside these transformative technologies of the past, investment in AI does not seem so out of scale, though AI has attracted capital at an unprecedented rate.
The chart shows the investments made in different technologies in present-day terms. Investments of $3 trillion were made during the railway mania of the 1840s and 1860s, telecom investments during dot-com boom years were at $1 trillion and AI investments are at $1.6 trillion from 2013-2024.
But even when a technology proves to be longstanding, investors eager to get in early can lose out if they make a wrong bet, even as others make it big. It’s too soon to say how the AI boom will end and who will be the ultimate winners and losers — for now there’s no end in sight.
Estimates for the costs of the Manhattan Project and the Apollo Program come from the National Park Service and The Planetary Society, respectively. Estimates for investments in AI come from Stanford University HAI AI index report 2025. Though AI existed before 2013, Stanford’s data on investments in this wave of the technology begins in that year. Estimated investment amounts in telecom networks from 1996 to 2000 from the Brookings Institution adjusted to 2024 inflation numbers. Estimates for the railroad boom investments, also in 2024 dollars, are from research by Andrew Odlyzko, a mathematics professor at the University of Minnesota. Company valuation and revenue numbers come from LSEG data for Microsoft and Nvidia and from media reports for OpenAI. Valuation estimates for Nvidia and Microsoft in 2022 as of the end of the third quarter. In 2025, they are as of November 30, 2025. For OpenAI, which is not a publicly listed company, the valuations are as of the most recent media reports to those dates. For Nvidia and Microsoft, revenue estimates in both 2022 and 2025 are the sum of revenue in the four quarters ending with the third quarter of the respective year. For OpenAI, revenue estimates are for all of 2022 but only the first half of 2025, per media reports.
Ella Koeze, Lisa Shumaker
