Our old friends, fear and greed, loom large here. Clearly, greed is uppermost to the extent AI is perceived as the next gold rush. If new AI IPOs result in quick doubles or more, who would not want to participate in such a wealth-creation event? Meanwhile, fear manifests in two ways. Most investors with significant allocation to stocks as an asset class naturally fear a crash in the stock market, especially after the kind of upward moves we’ve seen in the past year or two. But there is also so-called FOMO, or the fear of missing out on some hot innovation that promises untold riches down the line. AI seems to epitomize both varieties of fear (although you could argue FOMO is just another manifestation of greed).
From where I sit, well into registered retirement income fund (RRIF) age, AI is a theme that young investors have little choice but to embrace, at least in part. Growth is the preferred strategy for those just starting out on their investing careers, particularly for tax-free savings accounts (TFSAs). And if any transformative innovation seems poised for major growth in the long term, it seems to be AI.
But for those in the retirement risk zone—perhaps including you, dear reader!—there is potential danger in jumping whole hog onto the AI bandwagon. While AI should not be ignored as a key growth play for the ”satellite” or “explore” portion of a portfolio, I’d be wary of building the core of a well-diversified global portfolio around AI.
The best online brokers, ranked and compared
All these trade-offs were explored in a webinar held early in July by my site in conjunction with The Successful Investor/TSI Network.ca, which focused on the prudent way for (relatively) conservative investors to play the AI theme. While I did announce the webinar on my site, this column is my first writeup of the event.
In the webinar, TSI account executive Bob Wiseman began by noting that in the first quarter of 2025, investors “quickly embraced” AI, pouring into high-growth AI-driven stocks like Nvidia, Amazon, Microsoft, Alphabet, and a host of AI start-ups. He didn’t name individual stock names of the latter group but based on my own dabbling in the sector, they would include “neo-cloud” providers like Coreweave and Nebius.
AI sector whipsawed by events
However, by the third quarter of 2025, fears of a possible AI bubble materialized, either in some investors’ minds or in media coverage of the phenomenon. Investors, Wiseman said, started to ask themselves whether AI costs would prove to outstrip revenues for years or even decades. That mindset then gave way to a “flight to reality: a retreat from AI-linked stocks to old-economy sectors like resources, manufacturing and utilities.”
Wiseman described several subsequent events that either drove the AI theme higher or lower, such as China’s DeepSeek scare, which made names like Nvidia plunge in January 2025, and an MIT study in August 2025 that found 95% of companies saw no returns from generative AI. But markets also rose in May 2025 when Nvidia chip sales to data centres surged after markets reacted from the initial tariff shock. And there was a general jump in markets in February 2026 after payment processor Block announced AI had helped it to axe 4,000 jobs.
In short, this is a volatile sector. As some unknown wag once said about predicting future directions of the stock market, “it will fluctuate.”
X
Even so, Wiseman said, AI continues to matter because it is having a “transformational impact” on such varied major economic sectors as health care, manufacturing, financial services, advertising, software development and media/entertainment.
I was struck by one slide that showed conservative investors may be better off sticking with the big names that leverage AI, including the so-called “hyperscalers,” rather than the lesser known pure-play AI startups. It showed that the one-year performance of the Global X Artificial Intelligence & Technology ETF (AIQ/Nasdaq) was relatively flat, while household names like Intel were up more than 500%. Similarly, AIQ was trounced by such long-established companies as Alphabet and Cisco Systems. During the period, Alphabet was up more than 100%, versus 47% for AIQ, and Cisco was up almost 83%. (Disclosure, I don’t personally own AIQ but do have a small speculative position in a similar ETF: the VistaShares Artificial Intelligence Supercycle ETF [AIS/Nasdaq].)
A more cautious approach to investing in AI
The Successful Investor’s approach to investors partaking in the AI revolution is suitably cautious. Enthusiastic investors may see new ideas sparking market excitement and huge growth potential, the webinar warned, but they may also overlook the risks of an unexpectedly long wait for profitability or the fact that new innovations may disrupt existing businesses. As for the pure AI stocks and startups, some certainly promise major upside and some will succeed but history shows that “most will struggle or fail… as they always have in venture capital and junior stocks.”
Fortunately for conservative investors, there may be no need to experience FOMO about missing out on the IPOs of SpaceX or (soon), AI startups like OpenAI or Anthropic: “The biggest gains from AI will come from investing in established businesses that are already profitable and growing… They will profit by developing AI and its applications to their customers; that’s on top of applying AI to streamline their operations.”
Selling assets? Read our capital gains guide
All this serves as a textbook example of why seasoned investors may want to adhere to the three main principles long espoused by TSI founder and chief investment officer Patrick McKeough: invest primarily in established, dividend-paying companies; diversify across the five main economic sectors (manufacturing and industry, resources and commodities; consumer; finance and utilities); and avoid stocks in the broker/media limelight. The last goes double for “AI startups without a history of solid revenue, earnings, and preferably dividends.”
Wiseman says this approach has stood the test of time: “It has weathered and continues to weather AI volatility. While many of our AI stocks fell with the broader markets, they were in many cases the quickest to recover, with some moving up even as other ‘AI stocks’ fell on AI bubble fears.”
The AI theme in perspective
I’ve certainly adhered to this philosophy myself. After all, Pat McKeough was once my personal financial advisor. Even so, I do dabble in some lesser-known AI stocks that don’t make TSI’s screen for their pooled funds or discretionary investment management services. For the most part, though, I prefer to get initial exposure to AI through specialized sector ETFs. Full disclosure: in addition to owning AIS, as mentioned above, I have a tiny position in the Roundhill Memory ETF (DRAM), which owns most of the surging memory stocks like Micron, SE Hynix, Samsung, Sandisk, and many more.
To me, AI is just one sector, so it deserves no more or less weight than the five economic sectors specified by McKeough. Perhaps it should constitute its own sixth sector, depending on your age and risk tolerance. In this respect, it’s a bit like my similarly cautious approach to crypto investments, which I believe should not exceed 5% of a total portfolio. I’d prefer closer to 1% or 2%, if that. But as I would do if I chose to take a flyer on an individual AI startup, I keep the positions relatively tiny and would be quick to sell half on any double in order to “play with the house’s money.”
