Underwriting the Future of AI - How Trillions in AI Funding Could Reshape the Global Economy

Big Tech is in the midst of the largest capital spending binge in corporate history. From 2025 to 2028, $2.9 trillion is estimated to be spent on chips, servers, and data-center infrastructure across the AI ecosystem.² Google, Microsoft, Amazon, and Meta Platforms are set to spend nearly $400 billion this year on AI infrastructure. OpenAI alone has committed to $1 trillion in future AI-related spending.³ Wall Street has given Big Tech their blessing thus far as stocks have reached new highs, including Nvidia becoming the first company to surpass $5 trillion in market capitalization this year. However, beneath the euphoria is the harsh mathematical reality of AI funding. The financing and revenues needed for projected AI spending do not exist yet.

Big Tech Capital Expenditures, Quarterly

Source: WSJ. October 2025. “Big Tech Is Spending More Than Ever on AI and It’s Still Not Enough”

The scale of AI spending has reached unprecedented levels across the industry. The fundamental problem is that AI's compute demand is outpacing semiconductor efficiency, calling for dramatic increases in power supply on older grids. OpenAI has become synonymous with the AI trade and represents this spending well. Last year, Sam Altman started discussing trillion dollar plans for the buildout of AI infrastructure.⁴ Now, OpenAI is involved in deals for new datacenters that will take $750 billion in AI funding to build. Further, Altman has committed to spending $300 billion with Oracle over the next five years for cloud computing services. All in all, OpenAI has committed to $1 trillion in spending, or roughly 3.4% of U.S. GDP. So far, OpenAI only has large and growing losses to show for all of the spending, presenting a major AI funding gap. To cover the AI funding gap, roughly $100 billion will come from Nvidia in exchange for OpenAI equity, drawing comparisons to the circular financing deals in the dot-com era. Other partners such as Oracle and SoftBank will kick in another roughly $100 billion, but after their contributions there is still a $800 billion AI funding gap that will have to come from somewhere.

If the OpenAI funding gap is not covered, there could be a cascading effect in markets. When Oracle announced the cloud services deal with OpenAI, their stock soared 36% as the company gained $248 billion in market value.⁵ If OpenAI cannot cash the checks that it wrote to Oracle, then Oracle’s stock would crater. Oracle is not alone in their AI funding exposure. Companies from Big Tech to land developers to power companies are highly exposed to OpenAI. The success of OpenAI is central to the success of the broader stock market, and their AI funding gap represents a major problem.

Capital expenditures as a percentage of sales, yearly

Source: WSJ. August 2025. “Big Tech's $400 Billion AI Spending Spree Just Got Wall Street's Blessing”"

There is currently a $1.5 trillion AI funding gap for planned AI infrastructure spending.⁶ For OpenAI to honor their $1 trillion in commitments, traditional financing routes will not provide enough AI funding. OpenAI booked a record $40 billion funding round earlier this year, but that will only be a fraction of the AI funding needed. Equity markets in total have raised roughly $600 billion in inflation adjusted funds through IPOs over the last 45 years. Private financing has raised a total of $800 billion over the last 4 years. Broader debt markets also fall short as even companies that are known for their indebtedness don’t have that kind of funding. Verizon, for example, only has a net debt of $164 billion. Further, Verizon’s debts are backed by $265 billion in assets and licenses, things that OpenAI does not have. If traditional financing cannot close the AI funding gap, the money must come from revenue which is no easier task. Companies that are known for their cash flows can’t generate cash to that scale. Apple, for example, generates roughly $80 billion a year. Currently, OpenAI generates large and growing losses.

To fund the AI spending spree, Bain estimates that $2 trillion in revenue will be needed by 2030.⁷ If estimated efficiency savings from corporations applying AI in their workflows are counted towards that number, $800 billion in revenue will still be needed to close the AI funding gap. Companies are beginning to see AI revenues pick up, but $800 billion in AI related revenues will be a high hurdle as most companies today remain stuck in AI experimentation mode. In order to reach the revenues needed to fund the AI spending spree, models and AI agents will need to improve significantly and companies will need to fully buy-in to their uses.

Concentrated investment in a single asset class and suppliers funding their customers has traditionally been an indicator of a speculative frenzy. OpenAI has committed to building 16 gigawatts of data centers, and Nvidia has committed to investing $100 billion in OpenAI to fund those data centers.⁸ For scale, the last two nuclear reactors built in the U.S. generate a total of 2.2 gigawatts, took 15 years to build, and cost $30 billion. To power OpenAI and Nvidia’s dreams, 15 of these reactors would be needed. For these kinds of aspirations, there remains a $1.5 trillion AI funding gap with no viable path to close it through traditional financing.⁹ Instead, the industry will need $2 trillion in new annual revenue by 2030. AI will undoubtedly transform the economy, but the question remains if revenues will grow fast enough to justify the pace of exorbitant spending today. Without new AI funding sources or massive revenue growth materializing faster than any technology in history, AI's multi-trillion dollar investments risk becoming the next dot-com bubble, only many multiples larger.

Sources:

1. Barron's. October 2025. “Bubble or Not, the AI Spending Binge Is Unprecedented in Every Way”
2. WSJ. August 2025. “Big Tech's $400 Billion AI Spending Spree Just Got Wall Street's Blessing”"
3. Barron's. October 2025. “Bubble or Not, the AI Spending Binge Is Unprecedented in Every Way”
4. Barron's. October 2025. “Bubble or Not, the AI Spending Binge Is Unprecedented in Every Way”
5. Barron's. October 2025. “Bubble or Not, the AI Spending Binge Is Unprecedented in Every Way”
6. WSJ. August 2025. “Big Tech's $400 Billion AI Spending Spree Just Got Wall Street's Blessing”"
7. Bain & Company. September 2025. “$2 trillion in new revenue needed to fund AI's scaling trend”
8. Barron's. October 2025. “Bubble or Not, the AI Spending Binge Is Unprecedented in Every Way”
9. WSJ. August 2025. “Big Tech's $400 Billion AI Spending Spree Just Got Wall Street's Blessing”