Big Tech Has $1.65 Trillion in Hidden AI Debt

Big Tech is carrying $1.65 trillion in hidden AI debt, according to a report. Treasury yields hit a 19-year high as tariffs and the Iran war escalate, with the 30-year Treasury yield reaching 5.18%, its highest since 2007. The U.S. national debt stands at $39.4 trillion, and interest costs are expected to be the second-biggest federal expenditure this fiscal year.

Big Tech Has $1.65 Trillion in Hidden AI Debt Plus: the truth about America's startup boom 67% Of American men believe they have an above-average sense of humor. Treasury yields hit 19-year high as tariffs and Iran escalate Big Tech is carrying $1.65 trillion in hidden AI debt Is AI really behind the rise of one-person companies? Trump’s New Tariffs and the Iran War Are Rattling the Bond Market Tariffs are back, and the conflict in Iran has intensified. Last week the Trump administration announced tariffs of between 10% and 12.5% on more than 80 countries. These duties replace temporary 10% tariffs imposed in February. Earlier in the week, Trump also imposed a new 25% tariff on exports from Brazil and announced a 50% tariff on Canadian goods, including wine, hockey sticks, and cement. Meanwhile, the war in the Middle East is intensifying. Last week, the U.S. military completed 13 straight nights of strikes in Iran, and Tehran-backed Houthi rebels attacked two Saudi Arabian tankers in the Red Sea. Concerns over an escalating conflict spiked oil prices to $100 per barrel, the highest level in eight weeks. Prices fell to $84 per barrel on Sunday as negotiations showed signs of progress. Tariffs and oil have been among the biggest contributors to inflation, and last week, the odds of a rate hike before 2027 surged to nearly 70%, up from 35% just two months ago. Treasury yields also surged. The 10-year surpassed 4.7%, its highest level since January 2025, and the 30-year Treasury yield hit 5.18%, its highest level in 19 years. The last time it was this high was 2007, right before the global financial crisis. Treasury yields reflect the interest rates that investors demand to hold U.S. government debt. The 10-year Treasury serves as a benchmark for lending across the economy — from mortgage rates to small business loans — and rises when investors expect higher inflation or stronger growth . The 30-year Treasury yield reflects how investors feel about the fiscal health of the U.S. government. When yields are higher, that can be a sign investors see U.S. debt as riskier to hold. Higher Treasury yields mean that the U.S. government has to pay more to service its $39.4 trillion in national debt. The Congressional Budget Office estimates https://www.cbo.gov/publication/62257 that every 0.1 percentage point rise in rates over the coming decade will increase interest costs by roughly $1.8 trillion. Interest on the national debt was the third-biggest expenditure in the federal budget https://www.pgpf.org/article/any-way-you-look-at-it-interest-costs-on-the-national-debt-will-soon-be-at-an-all-time-high/ for fiscal year 2025, costing nearly $1 trillion. That’s about $7,700 per American household. Interest payments are already expected to be the second-biggest https://fiscaldata.treasury.gov/americas-finance-guide/federal-spending/ expenditure for the current fiscal year. According to the bond market, this is one of the riskiest positions the United States has ever been in. Why? Is it AI concerns? Iran? Inflation? Probably all of it. But I think the biggest driver is the fact that we have almost $40 trillion in debt and that sum keeps growing because we are being led by an administration that appears to have lost its mind. You could make an argument that the investment in the Iran war will be the best investment of the last 50 years. Unfortunately, the entity registering the return on the investment is China. Fareed Zakaria made this case. China gained a Middle East that is turning away from the U.S. It gained a world more dependent on Chinese technology, especially clean energy: China controls 92% of global solar panel manufacturing capacity and 89% of lithium-ion battery capacity, and makes at least 70% of almost all clean energy technologies. America is hemorrhaging credibility, while China is building a reputation for being the adult in the room. A Pew survey released this month found that, for the first time in roughly 20 years, more people around the world view China more favorably than they view the U.S. Alphabet, Amazon, Meta, Microsoft, and Oracle Are Hiding $1.65 Trillion in Debt An investigation from Nikkei Asia revealed that Alphabet, Amazon, Meta, Microsoft, and Oracle are carrying $1.65 trillion in debt that does not appear on their balance sheets. That’s in addition to the roughly $1.35 trillion in debt they officially report. How is this possible? Instead of building data centers and buying semiconductor chips themselves, hyperscalers sign long-term lease agreements with data center operators. That allows them to offload risk and avoid recording the debt on their balance sheets. These data center operators are often shell companies that are created and funded by private credit providers like Blue Owl Capital, Apollo, and Blackstone. The private credit funds are financed in part by public pension funds and life insurance companies. For example, the California State Teachers’ Retirement System is one of the biggest investors in Blue Owl’s public fund. What does all this mean? It means that the entities calculating the ROI on building data centers aren’t Big Tech but the private credit funds. And, if the AI bubble bursts, the risk won’t fall on Big Tech — it will fall on the retirement savings and insurance policies of working-class people. Private credit isn’t regulated by the Securities and Exchange Commission or the Federal Reserve or other bank regulators. This means it’s harder to assess systemic risk. In January, a group of senators sent an open letter to Treasury Secretary Scott Bessent urging regulators to take a closer look at the private credit ecosystem. The senators wrote that the magnitude of AI investment “appears to drastically exceed realistic assumptions of business and consumer demand for AI products and services in the near-term” and that if AI companies are unable to service their huge debt loads, that could trigger “ a broader financial crisis that harms the economy.” However, the department responsible for investigating problems like this, the Office of Financial Research, created in response to the Great Financial Crisis, has had its staff cut in half and its budget cut by more than 40%. There’s reason to worry about private credit funds’ risk assessment. Blue Owl Capital reportedly agreed to invest up to $10 billion in future AI data center deals after a single 15-minute meeting with the developer. But, thanks to Trump’s deregulatory agenda, we won’t know if private credit is creating a systemic risk — unless something goes terribly wrong. It’s easy to say all leverage is bad and these private credit companies are way out over their skis. And, there’s some truth to the notion that we’re due for a recession and it will likely start, as most recessions do, in the credit markets. Having said that, one of the amazing things about America and why our GDP has grown 6x more than Europe’s since 2008 is that our private credit analysts and VCs are more comfortable with risk than the private credit analysts in Europe. Does that aggressiveness subject us to more potential downside? Yes. But on the whole, that risk aggressiveness, that willingness to lend against future cash flows is a feature — not a bug of the American system. Are We in the Midst of an Entrepreneurship Boom? Or Is Being a Founder Just a Fad On paper, America is in the midst of an entrepreneurship boom. According to the Census Bureau, Americans filed nearly 6 million applications last year to start new businesses — the most on record — and applications are up 15% year over year as of June. But that data doesn’t tell the whole story. The census separates new businesses into categories, and the ones most likely to hire paid employees are deemed “high-propensity.” Since 2025, high-propensity applications have remained flat, while one-person companies have surged 20%. This raises the question: Is AI enabling the rise of one-person companies, or has it just become cool to have a side gig? Entrepreneurship is “in.” LinkedIn reported a 69% year-over-year jump in people adding "founder" to their profiles in 2025. Gen Z is leading the charge: 51% of Gen Z adults have seriously considered starting a new business in 2026, compared with 27% of all Americans. Some point to this uptick in entrepreneurship as an encouraging economic sign. But the reality is that these businesses aren’t creating jobs. Compared with the mid-2000s, the number of “likely non-employer” businesses as a percentage of total new businesses has doubled. It’s not even leading to a rise in self-employment. According to the Bureau of Labor Statistics, the number of people who designate themselves as self-employed on the household survey is also down over the past 20 years. When students come to my office hours, they don’t want to talk about brand strategy. They want career advice. The most common thing I hear is, I have an offer from Google and JPMorgan, but I’m thinking about starting my own business. They come to me because I’m an entrepreneur who’s had some success, and they think I’m going to encourage them to start a business. My answer is universally: Don’t be a f cking idiot.Go to work for Google. Entrepreneurship is romanticized. I became a founder because I was too insecure to work at a corporation. Every time people went into a conference room, I thought they were talking about me. I resented people I thought were less smart than me making more money than me. I hated not understanding why decisions were being made without my input. But all of that is a small price to pay for what a great company offers — an unbelievable chance to build your skills, and the best way to get rich slowly in history. People like being able to call themselves a founder, because founders are cool, they go on podcasts, they have big social followings. But they’re not willing to take the actual risk it requires to build something that, 6 times out of 7, fails. So people end up in this limbo — I have my job, but I kind of want to be a founder — and that’s what I have an issue with.