AI reshapes debt, software and chip bets

Amazon has returned to debt markets with a large cross-Atlantic bond offering, seeking the equivalent of $37 billion to $42 billion in a mix of U.S. dollar and euro-denominated bonds. The issue follows a $15 billion U.S. dollar bond sale in November, the company’s first in three years.

The latest financing comes as analysts warn that hyperscalers’ borrowing for AI data centers is accelerating. Morgan Stanley has estimated that AI-related capital expenditures by companies such as Amazon, Meta and Microsoft between 2025 and 2028 could reach $2 trillion, with more than $1 trillion funded through new debt.

Amazon, which historically relied on internal cash to fund infrastructure, has guided for about $200 billion in capital expenditures this year, sharply above 2025 levels. The spending will focus on data centers, chips and cloud capacity tied to AI demand.

Operating cash flow is projected at roughly $140 billion in 2025, leaving a funding gap that the new bond sale is intended to help bridge. The company has said proceeds are for general corporate purposes, including AI-driven capex, acquisitions and potential share support.

Amazon’s long-term debt stands at about $65.6 billion, including the prior $15 billion issue. Adding the planned $37 billion to $42 billion would lift total debt toward or above $100 billion, though this is set against a market capitalization of almost $2.3 trillion and an investment-grade credit rating.

Analysts note that Amazon’s leverage metrics, such as debt-to-EBITDA and interest coverage, remain conservative, supported by rising operating cash flow as AWS growth accelerates. Free cash flow could turn negative temporarily in 2026 as AI investment peaks.

Other technology and cloud providers are facing similar funding strains. Oracle is planning $40 billion to $50 billion of capex this year to support its cloud infrastructure and its OpenAI partnership, and aims to raise $45 billion to $50 billion through a mix of debt and equity, about half via bonds early in 2026.

Smaller players such as Applied Digital and Coreweave have also tapped credit markets, contributing to investor concerns about contagion risk in tech. Bond spreads for tech issuers have widened modestly, and there are worries that sustained borrowing for AI could eventually pressure margins or force spending cuts.

While spending on AI infrastructure accelerates, some enterprise software providers are being reassessed for their exposure to AI-driven disruption. Rothschild & Co Redburn upgraded Intuit to Buy from Neutral, arguing its core products are among the most resilient to such threats.

Redburn raised its price target on Intuit to $700 from $670, implying about 46% potential upside. The brokerage expects Intuit to deliver around 13% annual revenue growth and roughly 15% annual free cash flow growth over the next five years.

Analysts highlighted that key Intuit applications, including QuickBooks and TurboTax, are underpinned by deep data sets, complex regulatory rules and network effects that are difficult for AI-first competitors to match. These features are seen as supporting pricing power and steady expansion.

Redburn said enterprise software valuations have fallen to historically low levels as investors weigh the risk that AI tools could undercut traditional platforms. It argued that many established vendors retain durable advantages due to embedded data and business logic, even if AI creates shorter-term pricing pressure.

The brokerage described an emerging “intelligence layer” built by frontier AI developers on top of existing enterprise systems. It said companies with unique data and complex operational logic, such as Intuit, are better placed within this framework.

Redburn also cited SAP SE and ServiceNow as software vendors enjoying strong protection from AI disruption, attributing this to deep enterprise data and integrated workflows that are hard for external AI tools to replicate. It nonetheless cautioned that AI could reduce switching costs between vendors in the near term.

According to the firm, share prices across the enterprise software sector may have over-corrected for AI risks. Its stance is that while disruption is a concern, entrenched providers with rich data assets and established customer relationships still occupy defensible positions.

In the AI hardware segment, investor sentiment toward Nvidia has become more mixed after a prolonged rally. Nvidia’s market value is now above $4 trillion, and its shares have gained around 1,300% over five years, leading some market participants to question further upside.

Despite the share price consolidation, Nvidia continues to post strong operating performance, with gross margins well above 75% and ongoing product innovation in AI chips. The company recently delivered a strong quarter that drew a muted market reaction.

Billionaire Leo KoGuan, known as a prominent Tesla bull, has disclosed buying roughly one million Nvidia shares, valued at about $180 million. He has rejected talk of an “AI bubble” and views recent volatility as creating a buying opportunity, and may continue adding to his position.

Commentary around Nvidia points to potential risks from rising competition in AI chips, possible shifts in AI demand and factors such as higher memory costs. At the same time, AI adoption is described as continuing, with secular tailwinds for Nvidia and peers seen as intact.