AI Bubble May Resemble 2008 Financial Crisis More Than Dot-Com Bubble? Bloomberg Warns: Worse Than Internet Bubble!

Artificial intelligence (AI) stocks have recently experienced sharp volatility, reigniting market concerns over an 'AI bubble.' Notably, however, the reference point for investor anxiety is quietly shifting. Bloomberg reports that while most people previously likened this AI surge to the 1999 internet bubble, an increasing number of voices now suggest the situation may more closely resemble the 2008 global financial crisis.

The report analyzes that AI and the internet bubble do share similarities: both involve emerging technologies with uncertain applications, and investors, driven by optimism, have aggressively bid up related stock prices to astronomical levels.

Yet, recent sentiment has changed, with more market participants beginning to question whether this comparison is flawed.

Vitaliy Katsenelson, CEO of the Investment Management Association (IMA), admitted that he previously described AI's rise as reminiscent of the 1999 internet bubble. But now, he says, this bubble appears to be continuously inflating, gradually evolving into something closer to the kind of bubble that triggered the 2008 financial crisis.

The AI Boom May Be 'Reassuring' Compared to the Internet Bubble

Bloomberg analyzes that if the situation truly follows the 2008 script, it would be a highly pessimistic signal. In contrast, the AI boom differs fundamentally from the internet bubble, offering some reassurance.

First, stock overvaluation accompanying technological innovation is a common phenomenon, as the value of groundbreaking inventions is inherently difficult to assess precisely. Past bubbles—whether in canals, railways, automobiles, or the internet—have all ultimately spawned new technologies that reshaped the economy. The trajectory of AI appears similar; despite sell-offs in the stock market, its expansion has not been halted.

Second, in terms of valuation, AI stocks are not as excessively overvalued as during the internet bubble. Current price increases are primarily supported by real profits generated from AI growth by companies such as chip manufacturers. As long as this profit trend continues, even further price increases may not necessarily be considered expensive.

In fact, tech giants Microsoft (MSFT-US) and Apple (AAPL-US), both of which weathered the 1999 bubble, now have valuations that are actually cheaper than they were at the time.

Moreover, most internet companies that went public around the turn of the millennium were deeply unprofitable, with many having no revenue at all. Their valuations were so absurd that even 'price-to-dream' ratios seemed unreasonable. In contrast, today's AI leaders like NVIDIA (NVDA-US) show outstanding profitability—there is a world of difference between the two.

Credit Bubble Concerns Emerge as a New Warning Sign

Nonetheless, the report points out that the market is increasingly focusing on a more troubling concern reminiscent of 2008: a credit bubble.

Looking back at the global financial crisis that devastated economies in 2008, the trigger was the subprime mortgage bubble, where borrowers ultimately could not repay their home loans, setting off a chain reaction.

Notably, compared to the internet bubble, the damage from a credit bubble collapse is far more severe.

When the internet bubble burst, the main impact was a shrinkage in investors' paper wealth. The subsequent 2000 stock market crash led to an economic recession, but by historical standards, it was relatively mild. Moreover, after the bubble, the internet continued to reshape society and the economy, ultimately benefiting countless people.

However, if a credit bubble bursts, it directly leads to losses, distressed asset sales, and even waves of corporate bankruptcies.

Katsenelson emphasizes that what truly crippled the economy 20 years ago was not the fall in housing prices itself, but the complete collapse of mortgage-linked financial products, which dragged down the banking sector and the entire financial system.

Oracle's Soaring CDS Sparks Debt Concerns

Bloomberg acknowledges that it may be premature to claim the AI data center construction boom will trigger a chain of landmines like in 2008. However, that financial crisis left behind a risk template worth heeding.

One notable signal is Oracle (ORCL-US) taking on massive debt to build data centers, causing the company's 5-year credit default swap (CDS) price to recently surge—exceeding even the levels seen during the 2008 financial crisis.

In other words, the most dangerous risk for AI is not that stock prices have risen too much, but the growing credit pressure forming between capital expenditures, debt financing, and free cash flow.

Jitesh Kumar, a credit strategist at Societe Generale, estimates that Oracle's implied debt default risk has already exceeded 16%.

The report notes that Oracle's situation may be an exception among hyperscale cloud providers, but the trend of ballooning industry-wide debt is already evident, giving the market ample reason to remain highly vigilant.

Debt Financing Is Rising, But the Interest Rate Environment Has Changed

The report also highlights another trend worth noting: this year, global hyperscale cloud providers are expected to raise over $200 billion in debt, significantly higher than last year's $125 billion—indicating that AI capital spending is increasingly reliant on debt financing.

For over a decade, this was not a market concern. After the financial crisis, the U.S. remained in a low-interest-rate environment, and during the pandemic, monetary policy remained loose, allowing tech giants to finance expansion, acquisitions, and data center infrastructure at low cost.

However, with interest rates remaining high, the financing environment is now fundamentally different from the past.

David Roberts, portfolio manager at Nedgroup Investments, points out that in the past, companies benefited from quantitative easing (QE) and low-cost post-crisis funding, enabling rapid expansion, competitor acquisitions, and business growth through borrowing. Now, as markets gradually return to a more normal interest rate environment, companies must bear higher borrowing costs.

For the AI industry, this change is particularly significant. If financing costs continue to rise while companies keep expanding AI capital expenditures, they must demonstrate revenue growth and cash flow returns more quickly. Otherwise, bond markets may not continue to provide abundant funding support.

Therefore, Bloomberg emphasizes that examining the AI investment boom from the perspective of credit market investors is actually a healthy and necessary exercise.

Compared to equity investors, who can more easily reference the historical precedent of the tech industry ultimately prevailing after the internet bubble, credit markets focus on whether companies have the ability to repay debt, whether cash flows are sufficient to support liabilities, whether future financing channels can remain open, and whether asset valuations have become excessively inflated.

The report argues that the next phase of market focus will no longer be just whether NVIDIA's stock has more upside, whether cloud providers continue to increase capital spending, or whether AI model capabilities continue to break through. Instead, the focus will shift to whether companies can generate sufficient free cash flow, maintain sustainable financing capabilities, meet expected data center utilization rates, and whether AI businesses can support increasingly high fixed assets and investment costs.

In other words, if AI reenacts the 1999 script, the market correction will primarily affect valuations. But if the scenario resembles 2008, what will truly be tested is corporate creditworthiness and financial health.

And Bloomberg believes that what the market fears most is whether the AI investment boom will gradually evolve from a valuation-driven rally into a comprehensive test of capital spending and the credit cycle.