Spending Over $1 Trillion Annually? Tech Giants' AI Investments Keep Growing, but Bond Markets No Longer Accept Them Unconditionally

The wave of generative AI continues to fuel a global investment boom in the technology sector, but the financial advantages of large tech companies are facing unprecedented challenges. As investments in data center construction, AI servers, and infrastructure rapidly expand, Alphabet (GOOG-US), Amazon (AMZN-US), Meta Platforms (META-US), and Oracle (ORCL-US) maintain strong revenue and operating cash flow growth. However, their future free cash flow (FCF) is shrinking rapidly and may even turn negative—indicating that the AI era is forcing tech firms to shift from a "light-asset" to a "heavy-asset" business model.

According to CNBC, citing Moody's Ratings data, global capital expenditures (CapEx) by major tech companies are expected to reach $785 billion in 2026 and could approach $1 trillion in 2027.

This massive investment demand is also changing tech companies' financing models. Moody's estimates that global hyperscale cloud providers will issue up to $175 billion in bonds this year—far exceeding the past five-year average of around $30 billion—showing that AI infrastructure development is increasingly reliant on external financing.

Shay Boloor, Chief Market Strategist at Futurum Equities, says the market has underestimated the fundamental changes AI is bringing to large tech companies' business models.

He notes that large tech firms were once seen as classic "light-asset platforms," where revenue growth far outpaced capital investment needs. But the AI era is changing this advantage. Future growth in software, advertising, and cloud services must now be built on massive data centers, power, and computing infrastructure, gradually transforming tech companies into hybrid entities combining software and heavy-asset infrastructure.

Bond markets are also beginning to reflect investor concerns.

Recently, credit spreads on bonds issued by major cloud service providers have widened from about 50 basis points (bps) over U.S. Treasuries two months ago to 78 bps—the largest increase since the tariff market volatility in April 2025—indicating that investors are now demanding higher risk premiums and no longer unconditionally accepting tech giants' debt.

Oracle has become the most watched case in the market.

After its free cash flow turned negative due to massive AI data center expansions, Oracle's credit default swap (CDS) spread has widened to over 125 bps, the highest level since the 2009 financial crisis.

Moreover, Oracle's 5.20% coupon bond issued in 2025 and maturing in 2035 had an initial credit spread of about 105 bps, but later widened to 175 bps, causing a clear decline in bond prices.

Amazon is also feeling pressure from the bond market.

According to Fortune, when Amazon issued $25 billion in corporate bonds in July, it had to offer an additional 18 to 21 bps in yield on long-term bonds to attract investors. The subscription multiple was about 2.5 times, lower than the 3.2 times seen in March.

Torsten Slok, Chief Economist at Apollo, also points out that the average subscription multiple for new bonds issued by large tech firms was around 5 times in February but dropped to about 2 times by July, indicating that bond investors are becoming increasingly cautious.

Stock markets are also beginning to question high AI spending.

Alphabet recently reported second-quarter revenue of $119.8 billion, a 24% year-on-year increase, with Google Cloud revenue surging 82%. However, the company simultaneously raised its 2026 CapEx forecast from $195 billion to $205 billion and warned that spending would remain high in 2027. Following this announcement, its stock price declined.

David Russell, Global Market Strategist at TradeStation, said that even if profits continue to grow, stock performance may not be sustainable if capital spending keeps consuming corporate cash.

He emphasized that the purpose of a company is to generate cash, not to keep increasing spending, and that the market is beginning to re-evaluate whether AI investments will deliver sufficient returns.

Moody's warns of deeper structural challenges.

The report states that Silicon Valley firms are gradually shifting from "light-asset models" to "heavy-asset models," and future competitiveness will depend not just on software, but on data centers, high-power AI servers, and long-term infrastructure.

Currently, the six major AI infrastructure firms tracked by Moody's—Microsoft (MSFT-US), Amazon (AMZN-US), Alphabet (GOOG-US), Meta Platforms (META-US), Oracle (ORCL-US), and CoreWeave (CRWV-US)—have accumulated direct debt of about $460 billion.

Even greater pressure comes from off-balance-sheet commitments.

Moody's statistics show that the total data center lease commitments of these six companies have reached about $1.2 trillion, with over $820 billion related to data center projects not yet in operation. Since companies must still pay long-term rent even if AI monetization is slower than expected, rating agencies now treat such lease commitments as equivalent to debt.

Moody's warns that the tech industry's shift from light-asset to heavy-asset models requires unprecedented levels of capital investment and financing, which could weaken corporate credit quality.

Currently, Microsoft and Alphabet still have strong balance sheets and financing flexibility. In contrast, Oracle, rated Baa2 with a negative outlook, and CoreWeave, rated Ba3, face relatively higher financial pressure.

Analysts point out that as data center lease commitments continue to grow, and AI model developers like OpenAI and Anthropic serve as both major cloud customers and investment targets, bond markets are changing how they evaluate tech giants. In the future, investors may no longer view large tech firms as pure software platforms but rather as infrastructure companies requiring massive capital support.