#Alphabet and Amazon Are So Desperate for AI Compute They're Raiding Foreign Bond Markets

TL;DR (Direct Answer): The AI infrastructure war has officially exceeded the capacity of traditional corporate cash reserves. In mid-May 2026, both Alphabet (Google's parent company) and Amazon made highly unusual moves: they aggressively tapped into overseas debt markets to fund their AI expansion. Alphabet disclosed plans for its first-ever yen-denominated bond sale, while Amazon prepared a massive debut offering in Swiss francs. Industry analysts project that Big Tech will spend an eye-watering $700 billion on AI infrastructure this year alone—a massive jump from $410 billion in 2025. This borrowing spree signifies a major shift: the hyperscalers are no longer just spending their massive profits; they are leveraging global debt to ensure they don't lose the race to Artificial General Intelligence (AGI).

#The $700 Billion Burn Rate

To understand why two of the richest companies on Earth are suddenly borrowing money in foreign currencies, you have to look at the sheer scale of the hardware they are buying.

The AI industry is currently in the "scale-up" phase. Training next-generation models like Gemini 4.0 or Anthropic's Claude requires hundreds of thousands of advanced GPUs (like Nvidia's Blackwell architecture) housed in gigawatt-scale, liquid-cooled data centers.

Wall Street estimates that total Big Tech capital expenditure (CapEx) on AI infrastructure will eclipse $700 billion in 2026. Even for companies like Alphabet and Amazon, which generate massive free cash flow from their core search and e-commerce businesses, writing checks that large requires strategic financial maneuvering. They are effectively draining their liquid cash reserves faster than they can replenish them, forcing them to turn to the bond markets.

#Why Yen and Swiss Francs?

If an American tech giant needs money, why not just issue standard U.S. dollar bonds? The answer lies in global interest rate arbitrage and investor diversification.

• Alphabet's Yen Play: Alphabet is preparing an issuance expected to total several hundred billion Japanese yen. While U.S. interest rates remain relatively stubborn, borrowing in yen allows Alphabet to take advantage of structurally lower yields in Japan. It also opens the door to a massive pool of Japanese institutional investors eager for high-grade, reliable corporate debt.
• Amazon's Swiss Franc Debut: Amazon has mandated major banks, including BNP Paribas and JPMorgan Chase, to structure a six-part debt offering in Swiss francs, with maturities ranging from three to 25 years. Switzerland offers a highly stable currency and deeply negative real interest rates relative to the rest of the world, providing Amazon with incredibly cheap capital.

By spreading their debt across different global currencies, these hyperscalers avoid flooding the U.S. bond market with too much tech debt at once, which would inevitably drive up their own borrowing costs.

#The Debt-Fueled AI Bubble?

This borrowing spree is raising alarms among some cautious institutional investors.

For the last decade, Big Tech was defined by its pristine balance sheets. Companies like Apple, Google, and Microsoft sat on literal mountains of cash, rarely needing to take on significant leverage.

The shift to debt-funded AI infrastructure introduces a new layer of systemic risk. If the promised revenue from AI agents and enterprise SaaS integrations doesn't materialize fast enough to cover the massive interest payments and depreciation of the hardware, these companies could find their margins violently compressed. However, Alphabet and Amazon's leadership teams view the alternative—losing the foundational AI race to Microsoft or OpenAI—as an existential threat worth borrowing billions to avoid.

#The Global Hardware Drain

This financial maneuver has a secondary, geopolitical consequence. By raising hundreds of billions of dollars globally to buy up the world's supply of advanced semiconductors and power transformers, American hyperscalers are effectively pricing everyone else out of the market.

Startups, academic institutions, and even sovereign nations trying to build their own sovereign clouds simply cannot compete with an entity like Amazon, which can casually raise billions in Swiss francs on a Tuesday just to buy more server racks. The AI infrastructure layer is rapidly consolidating into an oligopoly funded by global debt.

#Capability Stack: The 2026 Big Tech CapEx Boom

#FAQ

Are Alphabet and Amazon running out of money?
No, they are highly profitable. However, their available "free cash flow" is not enough to cover the astronomical $700 billion infrastructure build-out expected across the industry this year without severely cutting into their operations or halting stock buybacks. Debt provides a cheap bridge.

Why is AI infrastructure so much more expensive than traditional cloud computing?
Traditional cloud servers run on CPUs, which are relatively cheap and power-efficient. AI training and inference require massive clusters of GPUs, which cost tens of thousands of dollars each, consume exponentially more electricity, and require entirely new, expensive liquid-cooling facilities.

Is it safe for US companies to borrow in foreign currencies?
It carries foreign exchange (forex) risk. If the value of the yen or the Swiss franc suddenly spikes against the U.S. dollar, it will cost Alphabet and Amazon more dollars to pay back the debt. However, companies of this size use complex financial derivatives (currency swaps) to hedge against this exact risk.

Who is actually buying these tech bonds?
Massive institutional investors—pension funds, sovereign wealth funds, and global asset managers in Japan and Europe. They view companies like Alphabet and Amazon as virtually risk-free, making their bonds a highly attractive, safe place to park billions of dollars.