The AI Capital Expenditure Arms Race: Big Tech's Quest for Earnings Growth and Debt Sustainability

The artificial intelligence revolution has ignited a spending frenzy among Big Tech's hyperscalers, with capital expenditures (capex) projected to reach unprecedented levels in 2025. According to a Citigroup report, Microsoft, Amazon, Alphabet, and Oracle alone will spend up to $490 billion on AI infrastructure this year-surpassing earlier forecasts by $70 billion-driven by surging enterprise demand for generative AI tools. This spending spree, while a testament to the transformative potential of AI, raises critical questions about long-term earnings growth and debt sustainability.

The Scale of AI-Driven Capex

Big Tech's AI investments are not just large-they are strategic. Microsoft, for instance, plans to allocate $62 billion in 2025, with over half directed toward U.S.-based AI workloads to bolster Azure's cloud capabilities, according to a Bloomberg report. Amazon, under CEO Andy Jassy, is investing $100–120 billion, with AWS absorbing the lion's share to capitalize on what Jassy calls a "once-in-a-lifetime business opportunity," as CNBC reported. Alphabet's $75 billion commitment focuses on servers and data centers to support its Gemini AI model, while Meta is deploying $60–65 billion, including a 2-gigawatt data center and 1.3 million GPUs, according to a UBS analysis.

Oracle, meanwhile, has taken a more opaque approach. Its $300 billion OpenAI deal and reliance on finance leases-now totaling $4 billion in liabilities-suggest a capex-to-sales ratio of 58% when including off-balance-sheet financing, as a Morningstar analysis shows. These figures underscore a shift from traditional cash-flow-based investments to hybrid models that blend debt, leases, and equity.

Financing the AI Boom: Debt, Leases, and Risks

The financial architecture underpinning these investments is as bold as the spending itself. Microsoft's finance-lease liabilities surged 70% to $46.2 billion in fiscal 2025, as Morningstar reports, while Oracle raised $18 billion in bonds to fund its cloud expansion, the Citigroup report notes. Amazon and Alphabet, though less reliant on debt, are not immune to risks. Bloomberg Intelligence notes that combined capex for AWS, Azure, and Google Cloud could hit $200 billion in 2025, with Microsoft leading at $62 billion.

This debt-driven strategy, however, carries inherent dangers. Morgan Stanley analysts warn that an additional $2.9 trillion in AI infrastructure spending between 2025 and 2028 could strain balance sheets, particularly if returns on investment lag expectations, a point highlighted in UBS commentary reported by CFODive. The use of finance leases, while cleverly preserving free cash flow, also masks true capital intensity. For example, Oracle's $4 billion in lease liabilities represent just a fraction of its total AI infrastructure costs, the Morningstar analysis shows.

Earnings Growth: Promise vs. Peril

The long-term earnings potential of these investments hinges on two factors: monetization of AI services and operational efficiency. Microsoft's CFO, Amy Hood, has emphasized that cloud and AI demand "justifies increased capex," citing Azure's role in driving revenue through AI copilots and large language model (LLM) licensing, as CNBC reported. Similarly, Amazon's AWS division is expected to generate new revenue streams from AI-driven analytics and enterprise tools, according to the Bloomberg report.

Yet, the path to profitability is fraught. UBS analysts note that AI revenues are likely to lag capex for years, with monetization improving only in 2025 as enterprises adopt AI at scale, as CFODive summarizes. Worse, there are concerns about cannibalization: AI workloads could displace traditional cloud services, eroding margins. For instance, Meta's $60–65 billion investment in AI infrastructure may not translate to direct revenue unless it successfully monetizes its AI models through third-party licensing, UBS analysts warn.

Debt Sustainability: A Double-Edged Sword

The reliance on debt and leases to fund AI expansion raises red flags for investors. Oracle's capex-to-sales ratio of 58%-when including leases-suggests a precarious balance between growth and leverage, the Morningstar analysis argues. Microsoft's $46.2 billion in finance-lease liabilities, meanwhile, highlight the risks of off-balance-sheet obligations noted in the same analysis.

Citigroup analysts caution that the AI arms race mirrors past tech bubbles, with speculative behavior potentially leading to overinvestment, as the Citigroup report observes. Amazon's recent stock price dip following cautious AWS guidance illustrates market skepticism about the durability of AI-driven growth, as CNBC reported. For now, the jury is out on whether these bets will pay off.

Conclusion: A High-Stakes Gamble

Big Tech's AI-driven capex boom is a high-stakes gamble with the potential to redefine the tech landscape. While the scale of investment signals confidence in AI's transformative power, the reliance on debt and leases introduces significant risks. Investors must weigh the promise of long-term earnings growth against the specter of overleveraging and uncertain returns. As Morgan Stanley notes, AI could add 0.5% to U.S. economic growth between 2025 and 2028, a projection relayed in UBS commentary reported by CFODive, but only if these companies can convert their infrastructure bets into sustainable revenue streams.

For now, the market remains divided. Microsoft's stock rallied after its capex announcement, while Amazon's shares faltered, reflecting divergent views on the AI value proposition. The coming quarters will be critical in determining whether this spending spree translates into earnings growth-or becomes another cautionary tale of tech overreach.