All Eyes on Markets for AI Bubble Watch: Is It a Floater or a Popper?

1 hour ago 2

Analysis In an employee share sell-off this week, OpenAI achieved a nominal value of $500 billion. In terms of valuation, the posterchild of GenAI — which is yet to make a profit — left in its dust companies like Toyota, the world's largest automaker.

Non-believers might be interested in whether this combination of escalating asset value and burgeoning debt is anything to worry about...

To put this in perspective, Toyota [PDF] sells around 10.8 million vehicles per year, accrues an operating income (profit before tax) of around $32 billion and has a market cap of a mere $250 billion.

If this continues, it's not hard to imagine OpenAI might be worth double Toyota's value. Welcome to the world of the AI boom, where gravity-defying feats are possible so long as enough people believe they are.

There are certainly enough believers out there. GPU rental business CoreWeave told US financial regulators [PDF] on Thursday it had added an incremental $3 billion tranche of delayed draw term loans to buy certain equipment, hardware, infrastructure, and other systems.

In August, CFO Nitin Agrawal said the company had taken on a total of over $25 billion worth of debt and equity since the start of last year, all to fund the building of its AI cloud infrastructure.

The leverage of CoreWeave is a drop in the ocean of money need to make AI believers' dreams come true. Last week, management consulting firm Bain & Company predicted that funding the necessary infrastructure, would require the generation of $2 trillion in revenue by 2030.

According to Oracle, the former application and database company, there will be plenty of buyers for that infrastructure. It says it has a $455 billion spending pipeline from its AI datacenter customers. The problem is, some of them do not yet have the money. OpenAI is on the hook for $300 billion of Oracle's capacity, and the definitely-now-a-cloud-AI company needs to borrow at least around $100 billion over the next four years to build the datacenters required, according to KeyBanc's projections.

The fact that some of Oracle core customers don't yet have the money for all the spending against which it is borrowing huge sums has caught the eye of credit rating agency Moody's, which pointed out the significant "counterparty risk" in Oracle's projected growth – the possibility that another party fails to meet its obligations.

None of this deterred investors: Oracle's share price has more than doubled since the end of April.

Non-believers might be interested in whether this combination of escalating asset value and burgeoning debt is anything to worry about.

Writing in the Financial Times this week, US financial commentator Robert Armstrong was not yet sounding the alarm, but was careful to add a note of caution. He cites an argument from Dario Perkins of economic forecaster TS Lombard: that leveraged bubbles such as property markets are far deadlier than unleveraged bubbles in stock markets, for example.

"We hope this is right, but we are not totally confident. It is true that the big tech companies are mostly financing their AI investments out of free cash flow. At the same time, we hear a lot about private debt investment in data centres, and we have deep respect for the financial system's ability to conceal leverage in unexpected places, especially when it is in the grip of a ‘next big thing' narrative," Armstrong wrote.

Last week, Barclays Equity Research published a note which kicked the tires of an AI bubble scenario. It concluded that AI as an investment theme remains on a solid footing.

Although it is "frothy," it is not a bubble yet, the bank said. "Hyperscaler capex/sales is up to 25 percent, but still well below the 40 percent for major telcos during the dotcom bubble. Leverage is also a fraction of what it was then. Opex is rising as the AI talent war heats up, but is still being outpaced by revenue growth from hyperscalers' core businesses, with AI upside down the road as adoption accelerates."

Meanwhile, IT analyst giant Gartner has maintained there is not a bubble, although there will be an extinction among AI model builders, who will be bought by bigger players, or see their market taken by surviving players. The staggering amount of revenue required to pay for the infrastructure will come piecemeal, in tiny bits from cash changing hands through the tech we all use.

"It's going to be in every TV, it's going to be in every phone. It's going to be in your car, in your toaster, and in every streaming service. Every piece of technology that you consume today will have GenAI in it," John-David Lovelock, Distinguished VP Analyst at Gartner, told The Register last month.

But what are we actually doing with AI?

This week digital studio Xicoia unveiled AI "actress" Tilly Norwood, much to the outrage of Hollywood's flesh and blood workforce.

Another example of a debatable application of AI comes from an argument over whether it should become a stand-in nanny or storyteller. The UK's Guardian newspaper recounts one father using ChatGPT to answer his four-year-old's endless questions about Thomas the Tank Engine. "My son thinks ChatGPT is the coolest train-loving person in the world. The bar is set so high now I am never going to be able to compete with that," he confessed.

Hierarchical, dictatorial, and sometimes downright nasty, the world of Thomas the Tank Engine is sadly one familiar to us all. With fears of a recession already in the air, the US government is likely to enter a second week of shutdown. The question is, for how long can AI believers deny that our fate is in the fat controller's small hands? ®

Read Entire Article