The order book opened in London on Monday morning and was more than twice covered before the pricing call. Alphabet was selling EUR 6 billion, about $6.7 billion, of euro-denominated bonds, and investors had placed orders of just over EUR 17 billion, according to a memo from one of the banks arranging the sale that was seen by Reuters.
The deal runs across seven tranches, one floating-rate note and six fixed, with maturities from two to 19 years. Guidance on the shortest piece was around 60 basis points over midswaps; the longest sat at 205 to 210 basis points. Barclays, BNP Paribas and Deutsche Bank are joint bookrunners, according to a regulatory filing. Alphabet said it would use the proceeds for general corporate purposes, which may include repaying debt.
For a company that spent most of the past decade buying back stock, the second euro sale of the year is a signal about what the artificial intelligence buildout has done to its cash flow. Capital spending reached $51.36 billion in the third quarter, up 114% from a year earlier, and exceeded operating cash flow of $42.29 billion. Free cash flow was negative $9.07 billion. The company has guided to $195 billion to $205 billion of capital spending for the full year and told investors to expect materially more in 2027.
The shares fell about 13% after those results. Long-term debt, as compiled by Bloomberg, has climbed to roughly $125 billion.
The market Alphabet is tapping has changed in the past year, and Alphabet is part of the reason. Amazon’s EUR 14.5 billion debut in March set a record for the euro corporate bond market. Alphabet’s 3 billion Swiss franc issue was the largest corporate borrowing Switzerland has seen. European insurers, pension funds and bond managers, drawn by duration that matches their liabilities, have been the buyers, which is why the European Central Bank recorded international euro issuance rising about 30% in 2025 to more than $1.1 trillion, the highest since the currency was created.
That demand is not unconditional. The Nasdaq fell 5% in October as investors questioned both the scale of AI spending and the returns it will produce. Recent technology issues have cleared only after arrangers widened spreads. The bid is price-sensitive at the long end, where duration and doubt sit together.
The 19-year tranche is the interesting one. An investor buying it is underwriting Alphabet’s credit for longer than the current depreciation schedule on the chips and buildings the money will buy. Accelerators are depreciated over roughly five to six years, data centre shells over 15 to 30. A bond that matures in 2045 will still be outstanding when most of the hardware financed by this sale has been written off. That mismatch is not a defect in the structure; it is the structure. The question is what spread compensates for it.
Analysts said the pricing outcome matters less for Alphabet than for the cohort behind it. Microsoft, Meta, Amazon and Oracle have all moved into the European market or expanded their presence in it. Each new issue is a data point on where the marginal buyer of hyperscaler credit sets the clearing level, and the marginal buyer is no longer a technology specialist. It is a European insurance company comparing a 19-year Alphabet bond with a utility, a telecom or a national champion.
That comparison cuts both ways. Utility debt carries regulated returns and stable demand. Alphabet carries search advertising, cloud contracts and a capital programme whose returns are still being argued about. The credit is rated in the high investment-grade bracket, and the balance sheet can absorb several years of negative free cash flow. But the reason the company is borrowing in euros at 205 basis points over midswaps rather than funding from cash is that the cash is committed.
There is a second-order effect that rarely appears in the prospectus. When free cash flow is negative, debt stops being an optimisation and becomes a funding source. The distinction matters at the moment the cycle turns: a company that borrows to accelerate growth can stop borrowing. A company that borrows to cover a shortfall cannot. Nothing in Alphabet’s numbers suggests the second case yet. The first nine months of the year, however, moved the company closer to it than at any point since it listed.
For European credit investors, the appeal is straightforward. Alphabet is one of a handful of issuers that can absorb a EUR 6 billion print without a new-issue concession large enough to hurt the secondary market. The order book of EUR 17 billion tells you the accounts were there. What it does not tell you is whether they will be there at 180 basis points in 2028, when the same borrowers return with larger requests and a longer list of comparable deals behind them.
Alphabet declined to comment beyond the filing.
The bonds were expected to price later on Monday, and the final spread will be the number that matters. If the 19-year piece clears near the tight end of guidance, the market is saying that hyperscaler credit belongs in the same conversation as infrastructure. If it clears wide, or the book thins before the call, the message is different: investors will fund the buildout, but they want to be paid for the possibility that it does not earn what it costs.
Either way, the euro market now has a role in the American AI cycle that nobody assigned it. European household savings, channelled through insurers and pension funds, are financing data centres in Ohio and Texas. The arrangement is efficient. It is also new, and it has not been tested by a downturn.
