- HEADLINE & INTRO
Alphabet is no longer financing AI as if it were a product cycle – it’s financing it like a railway. By issuing an ultra-rare 100‑year bond in sterling, Google’s parent is effectively telling markets that AI infrastructure will matter for generations, not quarters. This isn’t just a clever financing trick; it’s a signal about how Big Tech sees its role in the global economy. In this column, we’ll unpack what this century bond really says about the AI arms race, who’s taking which risks, why Europe is suddenly a key funding source, and what all this means for the next decade of cloud and AI.
- THE NEWS IN BRIEF
According to the Financial Times, as reported by Ars Technica, Alphabet has mandated banks to sell a 100‑year bond as part of its first bond deal denominated in British pounds. The century bond will sit alongside a multi‑currency package that includes around $20 billion in US dollar bonds (increased from $15 billion due to strong investor demand) and a planned Swiss franc tranche.
The proceeds are aimed at financing Alphabet’s soaring spending on AI‑related infrastructure, particularly data centers and cloud capacity for its Gemini AI services. The FT notes that Big Tech and suppliers together are expected to pour close to $700 billion into AI infrastructure this year. Alphabet alone signaled up to $185 billion in capital expenditure for the year, roughly double the previous period, after annual revenues surpassed $400 billion. The company’s long‑term debt climbed to about $46.5 billion, while it still held over $120 billion in cash and equivalents at year‑end.
- WHY THIS MATTERS
A 100‑year bond is not just another line item in a funding plan; it’s a statement of identity. Governments, elite universities, and a few blue‑chip corporates issue century debt. When a tech company does it, markets are effectively treating it as quasi‑sovereign infrastructure.
Alphabet is using its balance sheet to shift part of the AI gamble onto investors who think in decades, not product cycles. Pension funds and life insurers that need ultra‑long assets to match their liabilities get what they want: long‑dated, investment‑grade paper. Alphabet gets what it wants: gigantic, fixed‑rate funding for data centers, subsea cables, and power contracts that will depreciate over decades.
It also exposes a less visible reality: even Big Tech cannot fund the current AI capex surge purely from cash flow without impacting buybacks, dividends, or flexibility for acquisitions. Selling 3‑year paper is easy; selling 100‑year paper is Alphabet saying, “We intend to be indispensable infrastructure in 2126 – and we are willing to be judged on that.”
Not everyone is buying the story. Some asset managers reportedly sat out the new issuance, citing low yields and worries about piling up exposure to hyperscalers whose obligations are increasingly tied to big, complex AI bets. For them, the risk isn’t Alphabet’s solvency today; it’s technological and regulatory uncertainty stretched over decades. The winners, for now, are Alphabet’s shareholders, who get leverage‑fuelled expansion without immediate equity dilution, and fixed‑income investors desperate for long duration in a world of scarce safe yield.
- THE BIGGER PICTURE
Alphabet’s move sits squarely in a broader shift: AI is turning cloud providers into the capital‑intensive giants telecoms operators once were. Oracle just raised $25 billion in one shot, with a flood of orders. Last year Alphabet itself placed a 50‑year dollar bond and raised billions in euros. Amazon and Meta are similarly hiking their capex plans to feed AI workloads.
We have been here before, in a sense. The dot‑com era saw massive over‑building of fiber networks, financed with cheap debt and optimism. Many investors were wiped out, but the physical infrastructure laid then underpins today’s internet. The difference now is that a handful of hyperscalers, not a motley group of telecom start‑ups, control the spending – and they are starting from fortress‑like balance sheets.
Compared with the late‑2010s ultra‑low‑rate period, today’s environment is harsher: interest rates are higher, and long‑dated funding is more expensive. The fact that there is still appetite for 50‑ and 100‑year tech paper says something uncomfortable about the global economy: investors trust Alphabet’s long‑term credit risk more than that of many countries.
It also shows how the AI race is re‑wiring capital markets. Instead of venture capital funding the frontier, it is the bond market underwriting the build‑out of gigantic, power‑hungry compute farms. Equity narratives drive valuations; bonds quietly pay for the concrete, transformers, and chips.
- THE EUROPEAN / REGIONAL ANGLE
The most interesting detail for Europe is not just that Alphabet is tapping sterling – it’s why. Issuing a 100‑year bond in pounds is reportedly cheaper than doing the same in dollars, and it opens the door to UK and continental European pension funds and insurers hungry for long‑term assets.
Europe, with its ageing population and large defined‑benefit pension systems, is structurally long liabilities. Under Solvency II and similar frameworks, insurers are encouraged to hold long‑dated, high‑quality bonds. Alphabet is designing a product exactly for that demand. In effect, European savers’ retirement money will help finance American AI infrastructure that may or may not be built on European soil.
This raises awkward questions. The EU is pushing for digital sovereignty, promoting European clouds and enforcing rules via the Digital Markets Act and the upcoming AI Act. Yet its deepest pools of capital are financing the very US hyperscalers Brussels is trying to curb.
For European cloud challengers – from OVHcloud in France to smaller regional providers – the bar just moved higher. Competing with players that can tap 100‑year funding at thin spreads is not just a technological challenge; it’s a funding one. At the same time, European governments and regulators will view this influx of ultra‑long‑term Big Tech capital warily, especially as data centers collide with Europe’s energy, climate, and land‑use constraints.
- LOOKING AHEAD
Expect this not to be a one‑off. If Alphabet’s century bond prices well and trades smoothly, Amazon, Microsoft, and even large chipmakers may start experimenting with longer maturities and niche currencies to finance AI infrastructure. The bond market loves a new benchmark, and a liquid 100‑year tech name in sterling could quickly become a reference point.
The key variables to watch over the next 12–24 months:
- Credit spreads: Do investors keep accepting very tight premia over government bonds, or do concerns about AI over‑investment widen spreads?
- Regulation: The EU AI Act, US competition policy, and energy regulation could all affect the profitability of AI platforms long before these bonds mature.
- Capex discipline: If AI monetisation disappoints – for example, if customers balk at the cost of advanced models – boards and rating agencies will push back on endless data‑center spending.
- Market structure: If a few hyperscalers entrench themselves as de facto AI utilities, century debt will look prescient. If the stack fragments or shifts to on‑prem and open models, today’s massive centralized builds may look excessive.
There is also a geopolitical angle: if US fiscal worries intensify and Treasuries cheapen, ultra‑long corporate debt may lose some of its relative appeal. Conversely, if sovereign risk in parts of the world rises, Alphabet’s bonds could look like a safe harbour.
- THE BOTTOM LINE
Alphabet’s 100‑year bond is a financial bet that AI and cloud infrastructure will be as enduring – and as systemically important – as railways, electricity grids, or telecom networks. For now, bond investors seem willing to underwrite that story at very slim yields. The real question for readers is simple: are we comfortable with a handful of private companies effectively becoming century‑scale infrastructure utilities, funded by our pensions but governed by their boards?
