“Profitability Limits and Bubble Warnings” – AI Empires Built on Debt Could Trigger a 2008-Style Financial Crisis

Big Tech Keeps Pouring Money into AI Despite Bubble Fears
Wall Street Turns to Bond-Hedging Strategies
Profitability Constraints Raise Systemic Risk Concerns

Even as talk of an artificial intelligence (AI) bubble grows louder, global Big Tech firms show no sign of slowing their AI investment spree. Driven by the fear of “falling behind in the AI race,” they are ramping up spending through massive bond issuances and loans — effectively “borrowing to invest.” While this strategy might pay off if economic returns materialize, a potential collapse of the AI bubble could trigger widespread defaults and ripple through the financial system. Some analysts warn that the funding structures of AI companies increasingly resemble those that fueled the 2008 global financial crisis during the subprime mortgage meltdown.

Surge in Big Tech Bond Issuances

According to Bloomberg and other outlets on November 9, Alphabet, Google’s parent company, plans to issue a total of $25 billion in bonds across the U.S. and European markets. The company will sell $6.5 billion worth of bonds in Europe and $17.5 billion in the U.S. The American tranches will span eight maturities, from three to fifty years, with the longest 50-year bonds yielding 1.07 percentage points above Treasuries. Investor demand has already reached $90 billion, more than three times the offer. Alphabet also issued $7 billion worth of euro-denominated bonds last April to raise funds.

Meta, the parent company of Facebook and Instagram, is also preparing a $30 billion bond sale led by Citigroup and Morgan Stanley, which has reportedly drawn $125 billion in orders. Another major tech firm, Oracle, issued $18 billion in bonds in September after signing multi-year cloud infrastructure contracts worth $300 billion with clients including OpenAI and Meta. The sale attracted nearly $88 billion in demand — multiple times the offering size.

Goldman Sachs reports that the total value of AI-related corporate bond issuances has surged this year. Bonds sold by Alphabet, Meta, and Oracle alone have reached $180 billion, accounting for more than one-fourth of all new U.S. corporate debt supply in 2024. AI companies require enormous capital to build and operate data centers, and Morgan Stanley estimates that global Big Techs will spend $3 trillion on AI infrastructure by 2028.

Rising Concerns over Bond Market Overheating

The sentiment in capital markets is increasingly cautious. As AI-related debt issuance balloons, investors fear that a bursting AI bubble could evolve into a systemic crisis. With hundreds of billions of dollars flowing into potentially unprofitable AI bonds, any default wave could reverberate across global financial markets.

According to Bloomberg data, capital expenditures by major Big Tech firms on AI infrastructure reached $320 billion this year — more than double the $151 billion invested in 2023 — and are projected to exceed $1 trillion by 2031. Venture capital (VC) investments in AI startups are also at record highs, surpassing $100 billion last year.

However, few AI services have yet demonstrated stable, monetizable business models. Many AI firms trade at inflated valuations, with price-to-earnings (P/E) ratios detached from fundamentals. For instance, data analytics company Palantir’s forward P/E exceeds 200 times. U.S. equity markets, heavily driven by AI stocks, are showing similar signs of overheating: the Shiller CAPE ratio has reached its highest level since the 2000 dot-com bubble, while the price-to-book ratio of S&P 500 firms now stands at 5.3 — above the 5.1 seen during the dot-com peak.

Profitability Barriers for AI Firms

Amid these mounting concerns, OpenAI’s Chief Financial Officer Sarah Friar further inflamed the debate. Speaking at a Wall Street Journal conference last week, Friar said she wanted to “create a new financial structure combining private equity, banks, and federal guarantees to fund AI infrastructure.” Her remark implied that OpenAI might seek backing from the U.S. federal government for its enormous AI investments.

Following backlash over the notion of using taxpayer-backed guarantees to fund corporate profit, Friar retracted her statement, and CEO Sam Altman later issued clarifications. Yet Wall Street’s unease centers on a different issue: despite years of heavy spending on infrastructure, OpenAI continues to burn cash without generating profits. The company’s valuation has soared to $500 billion, but it has never reported a single profitable quarter since its founding.

Whether these astronomical investments will eventually pay off remains uncertain. Wall Street estimates that global AI must generate at least $1 trillion in added economic value for companies to see meaningful returns. Consumer demand, however, has lagged far behind. OpenAI’s estimated annual revenue stands at $13 billion, less than 2% of Amazon’s projected $700 billion, and only about 5% of ChatGPT’s 800 million users pay for subscriptions.

Echoes of the Subprime Crisis

As AI companies race to secure liquidity, Wall Street’s investment banks, private equity firms, and lenders are becoming deeply entangled in their financing networks. Hyperscalers — the largest cloud operators — are diversifying funding sources through bonds, loans, and credit facilities, while brokerages scramble to hedge their growing exposure.

The most talked-about strategy in this environment is the Synthetic Risk Transfer (SRT) mechanism. The Financial Times recently reported that Deutsche Bank has provided multibillion-dollar loans for AI infrastructure and is using SRT structures to hedge default risks. Under this approach, the bank keeps the loans on its balance sheet but sells off the credit risk to outside investors — much like slicing and trading loan exposures.

Critics note that this architecture strongly resembles the Collateralized Debt Obligations (CDOs) and Credit Default Swaps (CDSs) that amplified the 2008 financial crisis. Like CDOs, SRTs bundle and repackage loan exposures, and like CDSs, they separate and trade credit risk independently. Designed to distribute risk, such instruments can instead magnify it during downturns. The parallel is all the more concerning given Deutsche Bank’s history — it was one of the key players that profited from selling CDSs tied to subprime mortgages before the 2008 collapse.

Adding to the anxiety, famed hedge fund manager Michael Burry — the protagonist of The Big Short who famously predicted the subprime meltdown — has reportedly bet against AI stocks. His firm, Scion Asset Management, purchased put options equivalent to one million shares of Nvidia and five million shares of Palantir as of the quarter ending September 30. Both Deutsche Bank’s SRT trades and Burry’s short bets underscore the same conviction: today’s AI bubble may be the modern echo of America’s pre-crisis housing boom.