Market Factors: Brace for turbulence, AI air pocket ahead

This edition of Market Factors starts with preparations for a correction in AI stocks and moves on to some credit concerns in private equity and data centre funding. We look at spy novels in the diversion and there’s quick hits as usual

Open this photo in gallery:

Technology

Buyable correction for AI stocks

One strategist believes that Tuesday’s gut-checking 7 per cent intraday drop by Nvidia Corp., the poster child for the top-performing market theme of data centre construction, is merely a sign of things to come. There are, however, significant differences between the dot-com bubble and the current AI stock rally that might make any serious sell-off a buying opportunity.

Nvidia NVDA-T stock fell from U$182.00 to U$169 in overnight trading Tuesday on news that Alphabet is in talks to sell its Tensor Processing Unit (TPU) chips to Meta Platforms Inc. META-Q. The fear was that the TPUs could displace sales of Nvidia’s Graphics Processing Units (GPUs) as the mammoth buildout of AI-supporting data centres ramps up. BofA Securities analyst Vivek Arya expects the market for data centres’ processing chips to quintuple to U$1.2-trillion by 2030.

Every 500,000 of TPU sales will add 3 per cent to 2027 earnings for Alphabet. I think Alphabet needs to take a leadership position in AI because their previously near-monopoly on search is probably going away. I use AI exclusively for search now.

The broader question concerns whether Nvidia has lost its aura of invincibility and investors will adopt a more skeptical, show me attitude toward the AI investment theme. This scrutiny could lead to lower multiples and stock prices.

Savita Subramanian is BofA Securities’ chief U.S. equity and quantitative strategist. Her 2026 outlook report, published on Wednesday, discussed a much deeper air pocket for AI stocks as almost an inevitability. At the same time, the current market is not as extreme as 2000s.

Under the heading “AI air pocket ahead, but this isn’t 2000” Ms. Subramanian writes, “recommended stock allocations are much lower than in 2000, earnings growth has supported [the rally], IPOs are smaller, speculation in unprofitable stocks is less extreme.” All of these factors suggest a market with reasonably sane investors that will step in and prevent a complete implosion of the tech sector.

The strategist concedes that overall market valuations are expensive. Nine of the 20 metrics she follows closely indicate a market more expensive than early 2000, just ahead of the multi-year bludgeoning that tech stocks subsequently endured.

Unbeknownst to Ms. Subramanian, BMO senior economist Sal Guatieri published a chart late Tuesday showing that U.S. market capitalization as a percentage of GDP (a Warren Buffett favourite indicator) has hit 172 per cent. This is well above the 2000 peak at just over 120 per cent and implies a lot of room for a market correction if the measure is headed back to the historical average at roughly 70 per cent.

From 1993 until now there has been a 20 per cent correction in the S&P 500 every three years and the last one started in December 2021. I would never suggest that “we’re due” as the sole reason to go defensive but valuations also imply a higher risk of a market downdraft next year.

There will continue to be questions about whether the AI hyperscalers will be able to recoup their gigantic investments in data centres. I think, however, that Ms. Subramanian’s argument that 2026 won’t be 2000 is compelling and I will be looking for bargains in the event of a major correction. Stocks that I’d be watching include Palo Alto Networks PANW-Q, Datadog Inc. DDOG-Q, Western Digital Corp. WDC-Q, Elastic N.V. ESTC-Q, and Credo Technology Group Holdings CRDO-Q.

Technology, Part II

AI credit shenanigans

Any time investors see the words ‘blue’ and ‘owl’ together they should stop and read further. Blue Owl Capital Inc. OWL-N focuses on private equity and data centre-related lending, two areas where the list of investor concerns continues to grow. Short interest on Blue Owl stock has climbed from 3.7 per cent to 8.9 per cent since the end of March 2024, frustrating management who continue to reassure shareholders that nothing is wrong.

There are no signs of stress in Blue Owl’s private equity funds but the lack of news is part of the problem. The value of holdings in private equity funds is not transparent and investors must trust internal valuations. To be fair, there are no reasons not to trust the integrity of Blue Owl’s valuations.

Data centres are a new initiative for Blue Owl. The company entered a joint venture with Meta Platforms to build a giant data centre in Louisiana with costs approaching U$10-billion.

Blue Owl management have repeatedly stated that the terms of their data centre deals assure strong returns. Investors continue to question the future returns on data centre investment, however, and this is also weighing on Blue Owl stock despite management pleas.

Blue Owl stock is likely to provide a sentiment gauge for private equity and data centre investment and is worth following for these reasons.

Diversions

A new (for me) espionage author

The Ringer’s Chris Ryan recommended a new spy novel author, Oliver Harris, and I’m in the middle of his book A Shadow Intelligence. The book has interesting elements – it’s based in Kazakhstan for one – and there are some complicated shenanigans. The jury (me) is still out but I think I’m interested enough to read another of his books no matter what.

The best spy novel of all time is, of course, John le Carré’s Tinker Tailor Soldier Spy and I’m unlikely to listen to any arguments on this front. Charles McCarry is probably my second favourite author despite his inconsistency. I’d avoid Shelley’s Heart, for instance, but the chronicles of the Christopher family are terrific.

Alan Furst also wrote some fantastic spy novels. The settings of his best books was the periphery of World War II where both Allied and Axis spies were busy vying for access to key commodities and other strategic advantages.

I would love to hear about other good spy novels from readers.

The essentials

Looking for our updates on market movers, analyst actions, stock technicals, insider trades and other daily, weekly and monthly insight? Click here to visit our Inside the Market page.

Globe Investor highlights

Central banks and ETFs have been getting all the attention when it comes to purchasers of gold. But Mike Dolan also points to a less-known one: Tether

It’s time for Telus to kick its addiction to dividend hikes, says Andrew Willis

Another warning sign for investors: U.S. tech companies’ share of S&P 500 earnings has been slipping even as their contribution to the index’s market value remains near multi-decade highs

How REIT managers are approaching the opportunities and risks from AI data centres

Quick hits

I keep thinking about Richard Bernstein’s investing rule, “Returns are best where capital is scarce” and whether it still applies when central banks turn on the monetary spigot every time the market drops 10 per cent.

How long has it been since a contrarian investment worked? Gold might qualify but I can’t remember the last time an investor went against the megacap tech tide and won big. Value investing is very much out of favour for the time being.

I keep reminding myself of the old market saying that the bear case always sounds smarter. Earnings have been good on both sides of the border, so maybe there’s no need to think of complicated reasons why stocks are going to drop.

See our full earnings and economic calendar here