Big Tech's Funds Drying Up, but "AI Investment to Continue Until 2028" [Weekend Money]

Nice Credit Rating's Online Seminar

Big Tech's Major Investments Lead to Weakened Cash Flow

Alphabet's "Financial Strength" Has Cushion, Amazon's Funds Depleted by Next Year

Although the rapid expansion of AI investment by big tech companies has resulted in negative cash flow, projections suggest that by considering external financing, they should be able to maintain their investment momentum. However, the ability of AI businesses to continue generating high returns was identified as a core variable.

Woong Kim, Senior Analyst at Nice Credit Rating, stated during an online seminar on September 22 that after examining the sustainability of AI investments at Alphabet, Amazon, Meta, Microsoft (MS), and Oracle, these findings were observed.

Capital Expenditure (Capex) and Free Cash Flow Trends of the Big Five Tech Companies. NICE Credit Rating

Kim noted, "Investment momentum for the big tech companies will be maintained through external financing based on their financial buffer capacity until 2028," adding, "Even in the expanding global data center market, these companies are expected to continue playing significant roles, and the financial conditions of big tech companies are not seen as critical obstacles to further AI investment."

According to a financial capacity analysis, all five of these big tech companies have experienced deteriorating cash flows due to aggressive capital expenditures in AI infrastructure. In the second quarter of this year, all companies except Meta and Microsoft reported negative free cash flow, showing a notable expansion of net cash outflows.

Furthermore, lease liabilities not yet reflected in cash flow and future purchase commitments were cited as factors expected to intensify future funding burdens for these companies. Uncommenced leases refer to data center lease agreements that have been signed but not yet activated and, therefore, are not recognized as liabilities on financial statements. According to Kim, "As of the end of June, the combined balance of confirmed contracts—including uncommenced leases and obligated purchase agreements held by the companies under analysis—stood at approximately 2.7 trillion dollars, which ranges from 1.4 times to as much as 7.2 times the companies' most recent capital expenditures." He explained, "From a cash flow resilience perspective, Oracle faces the greatest burden."

The investments of big tech firms are not merely financial burdens. The sharp growth in the cloud business has led to a rebound in overall revenue growth from 2024 onwards. Kim explained, "Previously, despite high revenue growth in the cloud division, its share of overall operations was too small to significantly contribute to company growth. However, following the full-fledged launch of AI investments, this structure has been reversed."

The order backlog for these companies surged starting in 2025. Analysis showed that, excluding Meta—which does not disclose order backlog—there are secured order backlogs representing between seven months to a maximum of two years' worth of total sales. Oracle, in particular, holds an order backlog equivalent to 9.5 years of revenue. If contracts proceed as planned, this ensures mid-to-long-term revenue visibility. Kim pointed out, "While order backlogs provide demand visibility, they also entail investment burdens. To fulfill these contracts, building new data centers, securing new leases, and advance procurement of key semiconductors are all essential."

Kim further noted that while AI investment has been successfully monetized, the pace of investment growth is accelerating even faster. He added, "The multiple of revenue-to-capital expenditure has recently fallen again, indicating that the efficiency of AI investment is declining. Even considering that non-operational assets that have not yet generated revenue account for 70-80% of annual investment depending on the company, the recent sharp drop in investment efficiency metrics shows a deepening time lag between investment and profitability."

"Investment Burden Can Be Managed Through Internal Cash and External Financing"

Nice Credit Rating assessed that the investment burden of big tech companies can be managed if internal cash resources are supplemented with external funding.

When evaluating their resilience based solely on internal funds, only Alphabet was found to have sufficient cash reserves to independently manage its significant investment costs. Amazon and Oracle are expected to deplete most of their internal funds by the 2027 fiscal year, making substantial external financing inevitable from the 2028 fiscal year. Amazon is projected to require at least 94.1 billion dollars, and Oracle at least 55.7 billion dollars in external financing.

Kim went on to say, "For Meta, the financial burden will intensify from 2028 onward, potentially requiring a halt to shareholder returns or the active pursuit of external funding. For Microsoft, annual shareholder returns of 48.7 billion dollars are reflected in its cash flow, so suspending these payouts could free up considerable additional capital."

While issuing corporate bonds or increasing external borrowing would raise financial burdens, Kim predicted that these levels are manageable. Even if indebtedness is expanded, the major companies' debt ratios are expected to remain below 100% through 2028. He stated, "Our analysis indicates that each company’s net borrowing dependence will more than double by the 2028 fiscal year, resulting in a somewhat rapid increase in financial burdens. However, given their existing strong capital buffers and robust profit-generating capabilities, the increase in liabilities is not expected to undermine their overall financial structures."

Nice Credit Rating further stated that it will closely monitor whether AI developers—who account for a sizable portion of future big tech order backlogs—are able to withstand continued infrastructure costs, as well as trends in bond demand and the strength of external financing in offshore markets.

This content was produced with the assistance of AI translation services.