Is This 2008 Redux?

Executive Summary

Similarities of the AI Bubble to the Housing Bubble/Financial Crisis of 2008 include the excessive use of debt and complex financial arrangements leading to financial risks outside of traditional corporate and banking relationships. As written about in previous newsletters, AI demand for investment in data centers is resulting in complex financing arrangements – often financed through private credit, institutional lending, and off-balance sheet structures (see graph). If demand falls short of optimistic expectations, the major AI firms are still on the hook for their lease or service agreement payments. Now Congress is debating allowing 401(k) plans to invest in private equity and credit. Members of such plans then become the “unknowing” holders of this debt. The central lesson from 2008 is that financial crises often emerge not from the underlying economic opportunity itself, but from the excessive leverage and complex financing structures built around expectations of its future success. And finally, the AI numbers dwarf the Financial Crisis numbers.

For further analysis, continue to read The Details below for more information.

“The panic of 2007 was a result of the transformation of traditional banking into an off-balance sheet and derivatives world.”
–Gary Gorton, Economist and Professor at Yale School of Management (2008)

The Details

Recently, there was an article in The New York Times entitled, “This Boom Looks a Lot Like the Housing Bubble,” in which the author, Jen Harris highlights some of the similarities of the AI (Artificial Intelligence) Bubble with the machinations of the Real Estate Bubble and Financial Crisis of 2008. The Financial Crisis resulted in roughly $700 billion in bailouts – in today’s dollars – and pales in comparison to that which occurred during the pandemic, and what could result from the AI boom. The more disturbing effect from the Financial Crisis was the loss of almost nine million jobs.

The rapid expansion of AI is creating enormous demand for investment in data centers, GPUs, computing infrastructure and electricity generation. Increasingly, this investment is being financed through private credit, institutional lending and off-balance-sheet structures involving special purpose vehicles (SPVs). Excessive leverage, complex financing arrangements and the potential for financial risks to accumulate outside traditional corporate and banking balance sheets parallel the Financial Crisis. SPVs are being heavily utilized by the AI giants to keep the assets and their related debt off of their balance sheets. A long-term service agreement or lease is used to fund the SPV enough to pay the current debt obligation. But what happens if the demand projected never materializes? As Harris pointed out in her article, the Financial Times reported that the projected annual revenue at OpenAI was 28% lower in September than was previously expected. And The Wall Street Journal suggested that revenue growth at Anthropic, OpenAI’s main rival, was softening. Harris stated, “A small group of companies are buying and selling large amounts of a hot product to one another, inflating demand for the product and obscuring the risks of the system in a web of complexity. […] Analysis by the asset manager Sona found at least 176 circular financing deals in the A.I. ecosystem that were worth $3.6 trillion over the past three and a half years.” 

The amount of this off-balance sheet financing has been estimated by some to be close to $1.65 trillion. However, Harris reported that a Wall Street Journal study discovered the top nine AI companies have amassed around $3 trillion in debt. If demand falls short of expectations, the major AI firms are still on the hook for their lease or service agreement payments. One might wonder who is financing all of this debt? This time the players have jumped on the loosely regulated private credit industry for financing. Is it any wonder why Congress is actively debating allowing participants in 401k Plans to invest in private equity and credit? Pension plans are already in the game. The plan is to offload the debt to the unknowing masses. Harris points out that the size of the current AI boom is greater than that of the Housing Bubble and a crash could wipe out more wealth than seen during the Financial Crisis.

“As bad as the financial crisis was, it began with a small slice of mortgages of a still-diversified economy. This time, the sector at risk is our economy’s growth engine. A.I.-related investment fueled an estimated one-third of the country’s recent economic growth, and the handful of companies most exposed to the A.I. build-out now account for roughly one-third of the entire S&P 500.” Everything now is dependent upon future demand for AI services and the ability to meet the demand. “Recent estimates indicate that A.I. would need to generate as much as $3 trillion annually in perpetuity to return the capital extended. That’s 9 percent of the U.S. gross domestic product, or twice what Americans spend on food.

Already, a closer look at the fundamentals of OpenAI and Anthropic, the most critical players in this bet, reveals a picture perhaps even worse than the headlines suggest. Right now, OpenAI expects to lose $280 billion by 2030 even assuming revenues grow tenfold over the same period. And both OpenAI and Anthropic have already slashed prices 40 percent to 50 percent from early summer, perhaps because China’s DeepSeek, the leading ‘open weight’ A.I. model, reportedly nearly matches their latest models’ performance at about 1.4 percent of the cost.”

The current AI financing boom shares several structural characteristics with the period preceding the 2008 Financial Crisis, particularly the growth of off-balance-sheet borrowing, increasing leverage, complex financing structures and limited transparency. My principal concern is that the financial system may be financing a long-term technological transformation through increasingly leveraged structures whose repayment depends on assumptions about future AI demand and asset values that have yet to be fully tested.

If those assumptions prove overly optimistic, losses could extend well beyond individual AI companies and private credit funds, potentially affecting banks, pension plans and wider corporate credit markets. The central lesson from 2008 is that financial crises often emerge not from the underlying economic opportunity itself, but from the excessive leverage and complex financing structures built around expectations of its future success.

Have the AI players gotten over their skis?

The S&P 500 Index closed at 7,580, down 0.6% for the week. The yield on the 10-year Treasury Note fell to 4.93%. Oil prices increased to $104 per barrel, and the national average price of gasoline according to AAA fell to $4.17 per gallon.

The long-term median valuation level for the S&P 500 using the Shiller P/E is about 16.13. The current reading is a whopping 150% above the long-term median. A drop in the S&P 500 of 60% would be required to bring its price down to the long-term norm. And of course, any drop below the median, resulting in an undervalued market, would require a drop of over 60%. These are not far-fetched numbers. A study of the aftermath of all of the prior great stock market bubbles reveals similar results. In fact, after the October 1929 crash, the Dow proceeded to drop around 89%! The Nasdaq 100 technology index fell about 83% after the Tech Bubble popped from 2000-2002.

These valuation methods are not timing indicators. They do not suggest a correction is imminent. What they do tell us is that the market will likely be net flat over the next decade, with a lot of volatility in the middle. Sustainable earnings growth can help lower valuation levels only if prices don’t jump in conjunction with the earnings increase.

The Tech Bubble was a great example of how using “forward” earnings in one’s valuation methodology is fright with risk. Using actual fundamental results, the stock market remains massively overvalued.

The S&P 500 Index closed at 7,624, down 0.5% for the week. The yield on the 10-year Treasury Note fell to 4.99%. Oil prices increased to $103 per barrel, and the national average price of gasoline according to AAA fell to $4.29 per gallon.