Where's the Cash?

Edward Chancellor (0:08)

Hello, and welcome to another episode of The Capital Cycle Podcast. This is Edward Chancellor, and I've got with me Charles Carter, who's a European Portfolio Manager with Marathon. Welcome, Charles.

Charles Carter (0:20)

Great to see you, Eddie.

Edward Chancellor (0:21)

So Charles, we're going to talk today about how investors are entranced by the AI bubble, which may or may not be entering its final stage.

But at the same time, there are plenty of investment opportunities in Europe that in your view have been neglected or are being neglected. And a two-tier market has opened up similar in many respects to the one that appeared during the TMT bubble of the late 90s and early 2000s, the division between the so-called new and old economy stocks. Back then, Marathon kept faith in the old economy stocks and eventually did very well by sticking to its guns. Now, anyone who's been paying attention recently will notice that the AI boom is being funded with complex and risky debt structures that you believe are typical of past booms. And your thesis is about following the money, which companies require money, and which companies are spitting out money. Classic capital cycle theme. But let's start with the AI investment boom at the moment.

Charles Carter (1:32)

Well, I think all technological revolutions require financial innovation to get them off the ground. And at risk of teaching grandmother to suck eggs, the railway boom in the 1840s was largely financed by the innovation of script, where you could basically just pay a small fraction of the capital up front, something like 10% of the share capital. And then later on in the TMT bubble, obviously, in the late 1990s, you saw the use of Vendor financing, various barter deals, a well-timed IPO, the use of tracking stocks, all of these innovations, which helped that investment in the information superhighway in the late 1990s.

Edward Chancellor (2:16)

But which was looking at our capital account book was roughly in the US about 500 billion in telecoms investment in that period. And now, right today, the investment is much bigger, both in nominal and real terms about money going into AI. And I think I read that roughly half of the ongoing trillions of dollars being invested in AI are debt, which is a lot of leverage given the relatively small if fast growing revenues.

Charles Carter (2:45)

There's not a lot of cash changing hands. So a lot of promises, IOUs, the use of options and vendor paper, all of which is redeemable at an indeterminate date in the future based on a vision of a transformed and highly profitable new world.

Edward Chancellor (3:02)

And while the AI world is based on promises of future cash flow that may never materialize, you believe there are investment opportunities that offer cash today rather than at some indeterminate stage of the future. As the title of your piece says, where's the cash?

Charles Carter (3:21)

Yes, and I think you talked at the beginning about a two-tier market opening up. And I think if you contrast what's going on in the AI world with, for example, what's going on in the very unfashionable corner of European smid, or small and medium-sized company investment, you're seeing there at the moment quite a lot of cash bids for companies, valuations which the public markets have lost all interest in defending. And these are businesses which generate cash from customers. They pay their workers, they pay their suppliers, they generate profits and cash flows, which they then return to shareholders.

And that sort of contrasts with the sort of the cashless society of AI built upon future promises.

Edward Chancellor (4:09)

On the topic of future promises in AI, you provide an example of rickety financial structures that support the AI data centers or Neo Clouds as they're called.

Charles Carter (4:21)

Yeah, so we came across this Australian company called Iren. It was a failed Bitcoin miner that has transformed itself into a Neo Cloud provider, and it's been able to secure something like $19 billion of funding over the last 12 months through the issuance of convertible notes, chipped-backed debt, customer prepayments and equity.

Edward Chancellor (4:47)

And NVIDIA, the chip jant, is at the centers of an AI financing spider's web, involved in vendor financing or what's called circular financing deals that are much bigger than anything we saw on the TMT bubble. It's also involved with Iren.

Charles Carter (5:06)

Yeah, that's right. They did a deal back in May, which involved a five-year contract worth $3.4 billion to rent capacity from Iren for NVIDIA's own workloads. And at the same time, it took up rights to acquire $2.1 billion of Iren stock at a premium valuation, contingent on the delivery of 600,000 NVIDIA chips to the company.