« Mid-Morning Art Thread | Main

Wednesday Morning Rant

Betting the Business

The AI mania continues to impress, not just due to specific mind-boggling decisions but the wider market environment. There is no shortage of madness in either, and they can help to explain each other. My favorite recent example involves the Japanese investment firm SoftBank and its pet company, OpenAI. SoftBank has gone all-in on AI, and OpenAI receives much of its investment. SoftBank has invested heavily in OpenAI and its Stargate Project, which aims to invest half a trillion dollars over a few years.

And SoftBank is not just blowing smoke. The money is actually involved, and not a small amount of it. There has been one story in the SoftBank/OpenAI financing saga that has been unfolding over the past few months, and it - and the market's response to it - has been fascinating. Back in March, SoftBank announced that it had secured short-term financing to the tune of $40 billion to dump into its AI investments, with OpenAI getting 75% of that. It went onto SoftBank's corporate card, but as anyone who has ever financed a big-ticket item with a credit card knows, that's a bad way to do it.

And it became immediately clear that the lenders were balking over risks as SoftBank worked to refinance the debt. Just a couple of months later in June, investment analysts noticed that SoftBank is really exposed to AI, to the point that the investment firm is, in some ways, becoming a proxy for OpenAI itself, just with compound leverage. As it was put in the CNBC article:

... "Buying SoftBank is partly a leveraged bet on OpenAI," he said. "If OpenAI does well, leverage is great. But if OpenAI and other investments do poorly, the leverage will hurt SoftBank."

Whoa, whoa, whoa. Hold the phones. You mean that leverage amplifies both upside and downside?! The hell you say! This is as deep as sports commentary. "The team that puts up the most points will win the game, analysts say."

SoftBank isn't done. It has until next year to pay off that $40 billion credit card, and they have a plan for that:

SoftBank has said that borrowing would likely be ​repaid "through the utilization of existing assets and other financing measures."

In other words, it will try to borrow more favorably to pay off the credit card with a longer-term loan. It tried that in July, wanting to borrow ~$10 billion against its holdings in OpenAI, but couldn't get the terms it wanted because, get this, OpenAI is impossible to underwrite as it doesn't trade publicly and there's no way to recover value from it if there's a problem. If SoftBank can't pay, the lenders get nothing but non-liquid and possibly worthless shares in OpenAI - so lenders balked.

Flash forward another couple of months, and we arrive at last week. SoftBank has floated the largest non-financial bond offering in the region's history, offering an ~$11 billion bond, with $10 billion of it earmarked for paying off OpenAI-related financing. The company can't issue investment-grade debt (SoftBank is rated BB+, which is junk territory, and that is also where these bonds landed), already has a lot of debt (about $123 billion in net debt, though assets technically exceed that unless things unexpectedly go down in value...), about 1/3 of its value is a pure AI play (its holdings in OpenAI) and another third of its value is in a chip company (ARM) that is also riding high on the AI wave. Talk about exposure!

That's a great example of how speculative bubbles can evolve in one large company, but it gets better. After announcing this new bond the response to it was astonishing:

Shares of SoftBank Group jumped more than 7% on Thursday, as Japanese markets opened for the first time this week following a three-day holiday and the technology investor announced an $11.1 billion bond issuance to help fund its growing bet on OpenAI.

Borrowing stable money to pay off the credit card was good for a 7% pop in its share value. That is a mania, and it's everywhere. SoftBank serves merely as an example, because it isn't alone. It's going to be ugly when it blows.

But SoftBank is sure that this time, it won't blow. As its CEO Masayoshi Son put it:

... in a recent interview with CNBC [he], defended SoftBank's aggressive push into AI, calling the technology revolution "50x bigger" than the dot-com boom and arguing that any future correction in AI-related stocks would represent a buying opportunity rather than a structural threat.

50 times bigger? Let's hope not. And Son needs to hope that the "buying opportunity" doesn't work out like it did for Cisco Systems, which took 25 years to get back to the same nominal market cap it had at the height of the dot-com bubble.

A mania is no time to be wrong...