AI has taken over the stock market. bond market is next & more related News Here

The rapid proliferation of new technological infrastructure is usually accompanied by a boom in bond markets. The emergence of liquid markets for corporate credit on both sides of the Atlantic in the 19th century mirrored the pronounced growth of the railway industry. Before the advent of railway operators with their huge investment requirements – to buy land, lay tracks and ship goods – corporate finance was the province of small banks. Later it was linked first with national, then international bond markets.

The latest tech-infrastructure craze, this time over artificial intelligence, is similarly fueling the bond-market (see Charts 1 and 2). Since January, Meta, Nvidia and Oracle have launched $25 billion worth of individual bond offerings to finance their AI ambitions – none of them have ever raised that amount in equity. SpaceX has also added the same amount to the $86 billion raised by Elon Musk’s rocket-to-robot business from stock market investors last month. In March Amazon, whose computing cloud hosts many AI workloads, sold $37 billion of bonds. Alphabet issued a rare £1bn ($1.4bn) 100-year bond as part of a £5.5bn debt sale in Britain in February, along with more paper denominated in Swiss francs, days after it raised $20bn domestically.

Morgan Stanley expects there to be $350bn-400bn in AI-related investment-grade issuances in the US this year. That would account for about a fifth of the record $2.3trn in high-quality dollar-denominated bonds the bank forecasts companies will sell in the US. Another $50 billion could come from junk bonds tied to AI projects (out of total high-yield issuance of $440 billion). The Bank for International Settlements, the central bank of central banks, recently warned that AI projects may not make enough money to repay the debt that financed them. If bond investing is the “negative art” of choosing what to buy rather than what to buy, as it is sometimes described, then how much should bond investors avoid?

In some ways, the AI ​​bonus makes the corporate-bond market a little safer. The total debt of America’s five “hyperscale” cloud giants – Alphabet, Amazon, Meta, Microsoft and Oracle – rose by $228 billion in the six months to March. This is almost five times more than any such two-quarter increase in the past. The first four companies have made more profits for years than they know what to do with, so enjoy strong credit ratings. Microsoft’s debt is considered more secure than Uncle Sam’s. Only Oracle, the smallest and least profitable of the five, receives a moderate “B” grade from the major credit-rating agencies (although about half of U.S. corporate bond issuers get the same these days).

You might expect the credit-growth acceleration to increase spreads on corporate bonds, a measure of their riskiness relative to safer Treasuries. But because of the creditworthiness of the largest issuers, spreads remain on average around 0.8 percentage points, close to the lowest in a quarter of a century.

Investors are beginning to be more discerning between secured loans of hyperscalers and bonds funding some of the riskier projects that the tech giant is associated with. In April QTS Data Centers, a company owned by Blackstone, a giant manager of private assets, issued $4.6 billion of bonds to finance a giant data center in Georgia, of which Microsoft will be a client. The bond yield was initially 1.1 percentage points higher than an equivalent Microsoft loan. The spread has since increased to 1.6 points.

On the surface, risky debt issuance appears to be a great cause for concern. The wave of junk bonds, which started last May with a $2 billion bond offering by Coreview, which leases out top-tier chip capacity, is surging. Additionally, CoreWeave and fellow “neoclouds” like Nebius and Irene have issued billions of dollars in convertible bonds this year, which investors can swap for a set number of shares of the issuer.

However, the division between secured and risky debt can be confusing. Historically, financial volatility tends to form around assets that investors view as safe, rather than assets that they view as already illiquid.

When railway defaults increased after the financial panic of 1873, it led to a worldwide depression. Today no such cataclysm seems imminent. However, it is clear that avoiding the mess will be harder than ever. The AI ​​revolution is moving at such a fast pace that some early winners have already turned into losers, then, occasionally, back into winners – and vice versa. Predicting which companies will come out on top, and be able to cash in on their bonds when they mature in 10, 30 or, in Alphabet’s case, 100 years, will require more downside smarts than ever before.