At times the AI arms race seems like the waltzer at a fairground. As the ride spins faster and faster everyone is screaming, some through fear, some through excitement but no one can get off.

The amount of money that has been committed by companies as they seek to be the winners is astonishing. Google, Microsoft, Amazon and Meta (Facebook) are planning to spend $725bn on AI infrastructure in 2026, up +77% from 2025 and triple what was spent in 2024. This is before the investments they have made into AI companies directly.

But it is not the sheer volume of cash being spent, rather the circular structure of the financing that is worrying. The web of deals gets pretty complex and includes the incumbent hyperscalers mentioned above, the AI ‘start-ups’ that are now among the largest companies in the world (OpenAI, Anthropic), and the hardware companies, such as Nvidia. TS Lombard have calculated that as much as 85% of the AI ecosystem’s revenues comes from this capex recycling.

Such circular deals need not be a problem so long as the companies can continue to meet their commitments. Indeed, some in the AI industry say that this is the only way to finance a technological revolution at the speed with which it is being done. Although they would say that, wouldn’t they? Either way, it certainly allows a level of investment that outstrips the more traditional sources of equity and debt markets.

That’s not to say they haven’t accessed those traditional markets too. Google has raised $75bn in equity financing and $30bn in debt in recent months. Oracle has increased its debt by 50% to $156bn also. And into this environment we have possibly the three largest IPOs ever taking place in a matter of months: SpaceX, Anthropic and OpenAI.

SpaceX: To infinity and beyond?

SpaceX floated 5% of its shares a fortnight ago raising $86bn and leaving Elon Musk in control of 85% of the company valued at $1.8tn. At that point the shares were trading at more than 100 times last year’s sales. While space is infinite and SpaceX has significant first mover advantage in both its launch and satellite capabilities, that remains a punchy valuation.

Will Elon Musk one day repeat what Scott McNealy, CEO of Sun Microsystems, said in the wake of the dotcom crash to those investors who had purchased shares in his company at their peak?

“At 10 times revenues, to give you a 10-year payback, I have to pay you 100% of revenues for 10 straight years in dividends. That assumes I can get that by my shareholders. That assumes I have zero cost of goods sold, which is very hard for a computer company. That assumes zero expenses, which is really hard with 39,000 employees. That assumes I pay no taxes, which is very hard. And that assumes you pay no taxes on your dividends, which is kind of illegal. And that assumes with zero R&D for the next 10 years, I can maintain the current revenue run rate…Do you realize how ridiculous those basic assumptions are? You don’t need any transparency. You don’t need any footnotes. What were you thinking?”

Of course, it is true that in the technology sector valuations are much higher than other sectors of the stock market, justified by higher margins and impressive growth rates. But the circular nature of the sector’s financing reduces the margin of safety. What happens when one company delays their capex because of supply constraints? Most likely we would be looking at a correction rather than a bust. But one need only look back at the sub-prime crisis in 2008, or the collapse of Long-Term Capital Management in the late 1990s, to see that the knock-on effects can be brutal and unforeseen.

Investors in the Tenax Fund are looking to minimise the volatility of their returns. While we have some exposure to the AI trade, and even to SpaceX through our holding of RIT Capital Partners, at this juncture we are happy to look elsewhere for opportunities to invest in companies on more accommodating valuations but still offering strong growth prospects over the long term. And some of those opportunities may well be in the technology sector once the waltzer ride slows down and the market correction inevitably occurs.

* Sun Microsystems was a tech hardware company that designed and manufactured its own entire computer ecosystem, from chips, to workstations to software. Their shares peaked at a price of $64 in September 2000 (a multiple of 10x sales). The shares bottomed in October 2002 at $3. While the share price was boosted by a 1-for-4 stock split in 2007, the company was acquired by Oracle in 2010 for $9.50 per share – or $2.375 when adjusted for the stock split.

The above article has been prepared for investment professionals. Any other readers should note this content does not constitute advice or a solicitation to buy, sell, or hold any investment. We strongly recommend speaking to an investment adviser before taking any action based on the information contained in this article.

Please also note that the value of investments and the income you get from them may fall as well as rise, and there is no certainty that you will get back the amount of your original investment. You should also be aware that past performance may not be a reliable guide to future performance.