AI vortex
iStockphoto/natrot

Welcome to Soundbites, weekly insights on market trends and investment strategies, brought to you by Investment Executive and sponsored by Canada Life. For today’s Soundbites, we’re talking about AI investing with Corrado Tiralongo, chief investment officer with Canada Life. We talked about patterns of innovation, and we started by asking why AI has inspired such hyperbolic claims.

Corrado Tiralongo (CT): When assessing where we are in the AI investment cycle, the tone and behaviour of the market make it clear we’re in the euphoric stage. When executives describe AI as ‘more profound than fire or electricity,’ the language gives it away. The behaviour matches the rhetoric. Hyperscalers are racing to outspend each other on data centres, chips, and power. The fear of being under-invested is stronger than the fear of overspending. But importantly, this sits on top of strong fundamentals. Demand for compute and cloud services remains incredibly robust. Earnings are still strong. So, yes, it’s euphoria, but it’s grounded in real economic activity, which makes it an unusual moment.

On ‘the madness of herds’

CT: When we look at the herd behaviour in AI, the cycle follows a familiar psychological pattern: everyone moves together, investors, corporates, even governments. But the correction phase never happens as a herd. It happens gradually. One company hits physical constraints. Another realizes monetization is slower than expected. Investors notice margin pressure. The adjustment becomes a series of individual realizations. That’s why volatility has risen. The enthusiasm was collective. The reassessment is gradual.

Lessons from previous investment-return imbalances

CT: In comparing today’s AI cycle to past infrastructure booms like railroads and fibre optics, the lesson is remarkably consistent. The infrastructure succeeds. The first wave of investors often doesn’t. Railroads transformed economies but bankrupted early operators. Fibre-optics now power the internet, but their initial owners absorbed heavy losses. AI is following the same path. The investment is essential, data centres, chips, and power infrastructure, but monetization always lags. Innovation overshoots long before it pays off.

How earnings from U.S. hyperscalers show both strength and strain

CT: Looking at hyperscaler earnings this year, we see both undeniable strength and visible strain. AI cloud demand, chip sales and accelerating data-centre construction are all strong. The strain is also clear. Capital expenditure is rising even faster than revenue. Power and infrastructure constraints are becoming binding. And margins are tightening in certain divisions. It’s a moment where technological opportunity and physical limits are colliding. Both sides of the story are true at the same time.

Why fastest asset growth is not always best

CT: When comparing disciplined AI investors to those with runaway asset growth, the performance gap becomes clear. High returns attract capital, capacity expands faster than demand, and returns eventually fall. The companies that outperform pace their investment. They focus on measurable payback and protect free cash flow. In AI, that gap is widening. Some firms spend because they can. Others spend because they should. The difference matters. History shows that discipline beats scale when capital becomes abundant. The fastest asset growth often signals exuberance, not leadership.

Familiar investment patterns

CT: When we map the classic cycle of innovation — abundant capital, overshoot, and consolidation — onto AI, the parallels are unmistakable. Every major technological wave follows the same arc. Innovation sparks excitement. Capital floods in. Investment overshoots. Then the consolidation arrives. AI is deep into the abundant-capital stage. Overshoot is inevitable. It’s how costs fall, adoption spreads, and productivity gains eventually scale. The consolidation phase is where durable winners emerge. That’s the transition investors should be prepared for.

So, is AI a bubble?

CT: When advisors ask whether AI is a bubble, the most accurate answer is one of nuance. AI is not a full bubble yet, but it has unmistakable pockets of exuberance. Valuations have run ahead of the underlying cash-flow profile in several areas. Feedback loops between performance, media attention and capital flows are visible. Investor sensitivity to yields has increased, and the market is now reacting more sharply to signs of slowing momentum. But we’re not seeing the leverage, the indiscriminate speculation, or the financial excess that typically defines a full-blown bubble. What we have is a narrative bubble, not a balance-sheet one. The fundamentals still matter. Returns are still grounded in adoption, not just hope. For advisors, the message is balance: maintain exposure, but avoid concentration.

And finally, what’s the bottom line on AI investment?

CT: When thinking about the bottom line on AI investing, it’s helpful to separate the revolution from the investment cycle. The productivity gains will be real, but the path will be uneven because infrastructure always comes before monetization. The right approach is to maintain exposure but broaden it, from the headline infrastructure builders to the diversified beneficiaries in software, semiconductors and companies using AI to enhance returns on capital. The winners won’t be the firms that spend the most. They’ll be the ones that turn AI into sustainable earnings over time.

Well, those are today’s Soundbites, brought to you by Investment Executive and sponsored by Canada Life. Our thanks again to Corrado Tiralongo of Canada Life. Visit us at investmentexecutive.com, where you can sign up for our a.m. newsletter and never miss another Soundbite. Thanks for listening.

**

Go back to the article.