Vernimmen.com

Comment, question or quotation of the day

16-03-2026 : Amazon or the lesson in debt

On Tuesday, Amazon raised $37bn on the bond market, and on Wednesday €14.5bn. This is as much as the previous record for the largest bond issue by a corporate (Verizon, $49bn in 2013), which was used to finance an acquisition. 

In this case, the purpose of the financing is not external growth, but to fund capital expenditure that are growing almost exponentially:

$11bn in 2019

$31bn in 2020

$47bn in 2021

$72bn in 2024

$114bn in 2025

And $200bn announced in 2026, under the influence of AI and data centres, of course.

But given the size of its cash flow, its lack of dividend payments and share buybacks (only $5.7bn since 2013), Amazon was still in a position of negative net debt at the end of 2025: $105bn in cash against $87bn in bank and financial debt. With 2025 EBITDA of $13 bn, an AA rating and a market capitalisation of $2300bn, Amazon's debt capacity remains strong.

The structure of this week's record issue is the result of two constraints: sound financial management of the company and market absorption capacity. 

To reach the maximum number of investors, who have different constraints and objectives, and thus maximise the funds raised, Amazon has multiplied the options for its debt issuance: fixed rate and variable rate, which is rare for companies (it is more common for banks to borrow at variable rates); in dollars and euros; and over 14 different maturities of 2, 3, 4, 5, 6, 7, 8, 9, 10, 19, 20, 30, 40 and 50 years. A total of 19 different tranches, almost as many as the number of smartphones on its platform!

As any good corporate financier knows, by multiplying the number of tranches, provided that their unit volume subsequently ensures adequate liquidity on the secondary market, it is possible to avoid creating unnecessarily large repayment deadlines.  The largest will be in 2036 for $6 billion. And, except for 2028, the maturities of the euro-denominated bonds do not fall in the years when the dollar repayments are concentrated.

Taking on debt in euros for an American group with global operations, including a significant portion in Europe, does not expose it to exchange rate risk given Amazon's structural revenues in euros. It allows it to lower its apparent cost of debt, since debt contracted in euros is at a rate approximately 100 to 120 basis points lower than debt with the same maturity in dollars. As an example, while Amazon has taken on debt maturing in 2066 in US dollars at 5.95%, its interest rate in euros is 4.85% for a similar maturity (2064).

It is likely that Amazon will return to the bond market in the future to raise substantial sums, given its investment needs in the field of artificial intelligence.

14-02-2026 : Two recent developments on capital markets

Let's start with debt. Last week saw Alphabet, Google's parent company, issue £1 billion at 100 years and 6.125%, which is always an event, given how rare 100-year bonds are for companies. Rare, but not unheard of, since:

•    Motorola issued $100 million in 1997, 

•    JC Penney issued $100 million in 1999 (since gone bankrupt), 

•    EDF issued £1.35 billion at 6% and $1.18 billion at 6% in 2014, 

•    Oxford University in 2017 (£1 billion at 2.54%), and 

•    The Wellcome Trust in 2018 (£750 million at 2.5%).

As an investor, you need to have a strong stomach given the high sensitivity of this type of bond (16.8 for Alphabet bonds). This resulted in a 50% gain in six months between February and August 2016, or between July 2021 and September 2022, with the price of EDF's 100-year bond halving!

But this is just the tip of the iceberg! For several quarters now, the largest borrowers on both sides of the Atlantic, i.e. governments, have been reducing the maturity of the bonds they issue. In the United Kingdom, for example, the maturity of public debt is 8.8 years, its lowest level since 2000. In Germany, France and Italy, the average maturity will fall below 10 years this year for the first time since 2015.          

In the USA, this change is desired by the Treasury in order to reduce its budget deficit, since the cost of 2-year loans is 0.5% lower than that of 10-year loans. Of the $1.9 trillion deficit expected in 2026, this ultimately represents a fairly marginal saving of $9.5 billion for taking on liquidity risk, as the amount of debt to be refinanced through new issues automatically increases. Some will see this as a bet that rates will eventually fall, making future refinancing less expensive. Others will see it as a sign of investors' reduced appetite for long-term bonds issued by  government whose solvency is deteriorating.

On the equity market, what has been reduced is the length of the order book building phase during IPOs. Fluctuating between 8 and 10 days since 2013, it has been halved to an average of 5 days since the beginning of 2026. And as any average hides a degree of dispersion, the order book for the CSG defence group was closed in three days at the end of January. This obviously considerably limits the risk of a market event derailing the operation, for example because Mr Trump reiterates his desire to annex Canada, or because the Iranian dictator announces the closure of the Strait of Hormuz.

To achieve this objective, extensive pre-marketing work is carried out upstream with major investors, enabling us to gain a fairly clear picture of their interest, the price they might pay and their demand for securities before the official opening of the order book.