Borrowing 350 billion yuan in five years to bet on AI! The five major tech giants in the US are making history, or are they repeating Intel's mistakes?

In the past five years, in order to seize the high ground of artificial intelligence data centers, the five major tech giants in the United States have doubled their debt scale, paving the way for what they call "economic transition" investments through massive borrowing. Data compiled from the market shows that Alphabet (GOOGL.US), Amazon.com, Inc. (AMZN.US), Meta (META.US), Microsoft Corporation (MSFT.US), and Oracle Corporation (ORCL.US) - the five companies that have invested the most in new data centers in the United States - have accumulated approximately $350 billion in new debt over the past five years. These companies are convinced that cutting-edge AI services will bring in rolling income in the future. Investors have responded enthusiastically in the past, subscribing actively to various currency-denominated bonds issued by them. However, according to sources, Amazon.com, Inc.'s $25 billion bond offering this week faced a rare cold reception in the market, indicating that the willingness of the market to provide funding for investments in tech giants is not endless. For most companies, the cost of debt is still relatively manageable, and their profitability remains strong. Last year, the combined interest expenses of the five companies exceeded $10 billion, more than doubling from 2019, but still insignificant compared to the free cash flow of any one of them. As of the end of March, Alphabet Inc. Class C's free cash flow from operating activities after deducting capital expenditures amounted to $64 billion. However, some companies' balance sheets are already under pressure. Amazon.com, Inc. had negative free cash flow in the quarter ending March 31; Oracle Corporation's debt in 2025 is about 2.5 times its revenue, and cash consumption is expected to accelerate. S&P Global, Inc. downgraded Oracle Corporation to the lowest investment-grade rating on Thursday, citing the continued increase in AI spending. Software companies are typically high-margin businesses with low conventional capital expenditure requirements. However, for industry giants, the emergence of cloud computing has changed this landscape - server farms require massive investment. AI data centers are typically larger than previous facilities and the chips are more expensive, further increasing capital expenditures. DA Davidson & Co. analyst Gil Luria said, "The nature of these businesses is undergoing dramatic and sudden changes, which is why their cash flow is currently so weak." He believes that these companies argue that the expected return on investment from new AI services - especially relative to the low interest rates on their new debts - is sufficient to make expansion worthwhile. "That's what they told us," he said, "but you can see that investors are not at ease." In April, Amazon.com, Inc.'s CEO Andy Jassy expressed high confidence that "this will be monetized" and listed commitments from customers to its cloud business AWS for additional data center capacity. Meta's CEO Mark Zuckerberg recently stated in an interview that the demand for AI computing power continues to outstrip supply, "which gives us great confidence that continued infrastructure development will be a good investment." Equity investors are becoming increasingly cautious about how and when large cloud computing companies (often called "hyper-scale enterprises") will reap returns from their massive spending. Since the beginning of this year, only Alphabet's stock price has outperformed the S&P 500 index, while the stock prices of Microsoft Corporation and Oracle Corporation have fallen by over 20%. As these companies will release their quarterly financial reports starting later this month, bond market investors will closely monitor their spending plans - which have become a barometer of the health of the AI boom. Market concerns have shifted from "can companies keep up with industry expansion" to "how to finance" and "when will the returns come". Fitch Ratings corporate debt analyst Jason Pompeo said, "I'm not sure if we know whether Amazon.com, Inc., Alphabet Inc. Class C, Microsoft Corporation, and Meta will truly get investment returns. At this point, a lot of the demand hype seems to be largely a vision." The five hyper-scale enterprises have committed to capital expenditures of up to $725 billion this year, primarily for data centers and NVIDIA Corporation chips that power their operations. Sources of funding include their own cash, new loans, and some financing arranged off-balance sheet by companies like Meta. The continually rising debt burden may weaken the ability of companies to cope with future crises or technological changes. Even with decades of dominance in the tech industry, companies are not immune to the heavy pressure of debt - once their position weakens, risks will become apparent. When Pat Gelsinger took over as CEO of Intel Corporation in 2025, he stated that his top priority was to clean up the balance sheet to put an end to doubts about the company's survival. The company, which was still the world's largest chip maker until 2022, accumulated significant debt during the tenure of its predecessor to support investor returns, acquisitions, and ambitious capacity expansions. However, as the debt continued to accumulate, Intel Corporation failed to introduce competitive AI chips, missing out on the enormous opportunity that would have made NVIDIA Corporation the world's most valuable company. A wrong choice in manufacturing technology resulted in a loss of market share, declining revenue, and soon reported unimaginable losses. Wall Street expects the company to report losses for the third consecutive year in 2026. Ultimately, it was government relief and investments from NVIDIA Corporation - a company that had been suppressed by Intel Corporation for decades - that enabled the Silicon Valley old giant to stay afloat. DA Davidson analyst Luria said that the large-scale cloud companies are far from that point. Referring to the combined debt of these companies, he said, "It doesn't look bad. If they borrow another magnitude? That would be bad."