Amazon corporate branding displayed at a company facility

Amazon has raised £4.25 billion, or about $5.76 billion, in its first bond sale denominated in British pounds, expanding its sources of funding as the world's largest technology companies spend heavily on artificial-intelligence infrastructure.

The September 9 offering was split across four maturities, giving Amazon a mix of shorter and longer-term borrowing. Investor orders comfortably exceeded the amount sold, although demand was less intense than for Alphabet's record-setting sterling deal earlier in 2026. That comparison offers a useful snapshot of the market: investors still want high-quality technology debt, but they are becoming more selective as issuance accelerates.

Amazon's sterling bond sale at a glance

The company sold four tranches:

| Maturity | Amount issued | |---|---:| | 3 years | £1.25 billion | | 6 years | £1 billion | | 12 years | £1 billion | | 19 years | £1 billion |

Final investor demand reached approximately £10.65 billion, according to Reuters, down from around £12 billion before pricing. The issue was therefore roughly 2.5 times covered. Yields ranged from about 5.2% on the three-year notes to around 6.7% on the 19-year debt.

Those figures do not mean Amazon pays the same rate on every pound it borrows. Each maturity carries its own coupon and market yield, reflecting both the length of the loan and the premium investors demand over comparable UK government bonds. Longer debt generally carries more interest-rate and inflation risk, which helps explain the higher yield on the 19-year tranche.

Why Amazon borrowed in pounds

Issuing in sterling broadens Amazon's investor base. UK pension funds, insurers and bond managers often have pound-denominated liabilities and may prefer assets in the same currency. Amazon has already tapped euro, Swiss franc and Canadian dollar markets, so adding sterling reduces reliance on the US dollar market alone.

There can also be a practical currency match. Amazon earns revenue and pays expenses in the United Kingdom, and sterling debt can offset some exchange-rate exposure. Companies may use derivatives to swap the proceeds into another currency, however, so the denomination does not prove that all the money will be spent in Britain.

Amazon did not tie every pound to a named data-centre project in the reported deal. The proceeds can support general corporate purposes, refinance liabilities or fund investment. The broader context is nevertheless impossible to miss: Amazon Web Services is competing in a capital-intensive race to build data centres, acquire advanced chips, secure power and develop new AI services.

The technology debt boom is getting larger

Reuters reported that major hyperscalers have issued more than $200 billion of debt in 2026, already more than twice the amount sold during all of 2025. Hyperscalers are companies that operate enormous computing platforms and can add capacity across regions at exceptional scale. Amazon, Microsoft, Alphabet and Meta are among the best-known examples.

For years, the largest technology companies could fund most expansion from their huge cash flows. AI changes the timing. Data-centre campuses, networking equipment, power contracts and accelerators require large upfront commitments, while revenue from the new capacity arrives over many years. Bonds allow companies to spread that financing burden across time.

Debt is not automatically a sign of financial stress. Amazon has a large business and access to investment-grade markets. Borrowing can be cheaper than issuing new shares, and long maturities can lock in funding for assets expected to operate for decades. The risk comes from the collective scale: many companies are seeking money at once for similar projects whose eventual returns remain uncertain.

The European Central Bank has warned that heavy hyperscaler borrowing could crowd out other issuers and raise financing costs. If bond funds allocate more capital to technology giants, smaller or lower-rated companies may need to offer higher yields to attract the remaining demand.

How the deal compares with Alphabet's bond sale

Alphabet raised £5.5 billion in a five-part sterling transaction in February, including a rare 100-year bond. Reuters said orders for that offering were about five times the amount sold, compared with roughly 2.5 times for Amazon.

The gap does not by itself mean investors consider Amazon unsafe. Market conditions, deal size, pricing, available maturities and the volume of recent issuance all influence order books. But it supports the view that demand is not unlimited. Investors have more technology bonds to choose from and can demand better compensation when supply is abundant.

An order book can also overstate firm demand because investors sometimes submit larger requests expecting to receive only part of their allocation. The fall from £12 billion before pricing to £10.65 billion at the end shows that some orders were reduced or withdrawn as the final terms tightened.

What investors and businesses should watch next

Three indicators will determine whether the AI borrowing wave remains comfortable. The first is cash flow: cloud and advertising growth must keep generating enough money to service debt without restricting ordinary investment. The second is utilisation: new data centres need paying workloads rather than simply installed capacity. The third is power availability, because grid connections and long-term energy supply increasingly determine when AI infrastructure can go live.

Amazon's sterling debut suggests global investors remain willing to finance the build-out. It also establishes a pound-denominated yield curve that the company could return to in the future. For the wider market, the question is no longer whether technology companies can borrow. It is how much additional debt investors will absorb—and at what price—before the flood of AI-related issuance pushes yields materially higher.

The financing race is closely linked to the expansion of computing capacity and electricity supply. Read our report on Google's Finland AI investment and nuclear-power agreement for another example, or follow MatchUpWorld's latest Business coverage.