How the AI trend might correct violently over high costs

This is the second piece in a series exploring an outlook by Robert Almeida, Portfolio Manager and Global Investment Strategist at MFS Investment Management. Read part one of this series here.

Robert Almeida believes we are in a malfunctioning market environment. The Portfolio Manager and Global Investment Strategist at MFS Investment Management sees contemporary market volatility and its disconnection from underlying earnings metrics as a symptom of that malfunction. He argues that as the AI theme drives markets, we’ve entered into an environment where narrative trumps hard data and the reality of a high-cost, high-interest rate world that should benefit price makers has been overlooked in favour of AI promises that may be more muted than they claim to be.

Almeida argues that while equity markets have consistently moved higher in the life of this AI trend, there is huge day to day volatility under the surface. That volatility, he says, may be symptomatic of a late cycle market and indicative of a market that is adjusting its price assumptions. New data is overwhelming those assumptions, causing repricing. AI, he says, has sat at the core of this market given the sheer quantity of narrative that surrounds the theme.

“It’s been a bit of a 'something a CEO said on a podcast' environment.  You have database companies whose business models are built on an information asymmetry, a flow friction, and a CEO of an AI company says, ‘this is what models will do,’ and those stocks are down. Those database stocks are all down 30, 40, 50 per cent year to date. Yet if you look at their earnings, they’re up 15 per cent,” Almeida says. “But nothing’s going to change that narrative until those AI companies come public and then we can look at their income statements and we can see if they have zero revenues based upon data that they’re selling to a financial services firm. So I think ultimately you get volatility to the upside in these names that have been perhaps overly discounted, and you get volatility to the downside in businesses that it was assumed they were going to be able to provide commercial value in some of these verticals that maybe otherwise they cannot.”

Why AI costs could spark a market reset

Almeida’s view is that AI is being treated like “Ozempic for CFOs” where its capacity to automate tasks is allowing them to trim labour costs they couldn’t get rid of before. He argues, though, that in the United States businesses have spent more than a decade driving efficiency and cutting labour costs. Moreover, many of the promises of AI are now only being realized through human utilization and intervention. In short, many of these already lean businesses can’t cut much deeper as they still need skilled people with human judgement to use AI tools effectively. The trouble is that markets may be pricing in the idea of cheap AI that can replace more labour. If the reality turns out to be one where AI costs more than businesses anticipated and does less than markets anticipated, there could be a violent correction.

The costs of AI services could be at the core of this potential turn of events. Almeida notes that AI budgets are skyrocketing for many companies as they shift from traditional software as a service (SaaS) providers to AI tools. With SaaS models, companies pay for a per-seat subscription and their employees use the tools as much as they need to get the job done with the subscription cost staying the same. With tokenized AI tools, the cost of AI rises the more employees use it. Notably, Uber blew through its annual AI coding budget by April of this year due to employees’ heavy use of Anthropic’s Claude, according to Fortune. Even as companies find value in AI workflows, Almeida notes that there may be a convergence where traditional SaaS providers are still relied upon for core tasks that work with the AI tools and the human employee.

How advisors should be preparing clients now

For advisors seeking to protect against a possible correction in the AI theme, Almeida argues that markets are moving too fast for individuals to react in time. He believes that clients need to be prepared for volatility to come, noting recent significant spikes in the VIX and the risks that now come with the sheer volume of options being traded to satisfy retail demand for options strategies. If stocks see a significant correction and all those options are out of the money, dealers will need to sell that stock, which may compound that volatility.

“The future has no facts, and we’re going to get things wrong. And successful investing, to me, over the long term is about the magnitude of your mistakes, not the magnitude of your victories. And so I’ve positioned the portfolio in what I think are money-good assets, portfolios that own stocks of businesses that we think are going to be around for a long, long time,” Almeida says. “The value of advice is going to go up dramatically.”