Artificial intelligence has quickly become one of the biggest investment stories in modern markets.

Companies are spending enormous amounts of money on AI infrastructure. Technology companies are racing to develop increasingly powerful models. Data centres are being built around the world, while demand for AI chips, memory, networking equipment and cloud computing continues to rise.

And investors have rewarded many of the companies at the centre of this trend.

But there is an important question that investors should now be asking:

Has the market started to price in too much of the future?

Martin Becker, a senior market analyst at NorthernIndex.com who focuses on developments relevant to Canadian market participants, believes this is one of the most important questions facing investors today.

His concern is not that artificial intelligence is a passing fad.

Quite the opposite.

AI could transform industries, increase productivity and create entirely new businesses.

The concern is that investors may be paying prices today that assume almost everything goes right tomorrow.

"AI can be a genuine technological revolution and still create an investment bubble," Becker says. "The technology and the valuation are two different questions. Investors need to consider both."

A great technology does not automatically make a great investment

This is perhaps the most important point for investors to understand.

A company can have an excellent product, a strong competitive position and a huge addressable market and still be a poor investment if its shares are priced too aggressively.

History provides plenty of examples.

The internet changed the world. But that did not mean every internet stock purchased during the dot-com boom was a good investment.

Many businesses eventually disappeared.

Others survived and became enormously successful.

The problem was often not the technology.

It was the price investors were willing to pay for the promise of future growth.

The same distinction matters with artificial intelligence.

AI is already producing real economic benefits. Companies are using it for software development, research, customer service, data analysis, automation and many other tasks.

But investors are not buying AI technology itself.

They are buying companies.

And the price of those companies ultimately has to be supported by future earnings and cash flow.

AI valuations are becoming increasingly important

There are signs that investors should at least be cautious.

The Bank of Canada has warned that equity valuations remain elevated and that the market's concentration in large technology companies has increased. According to the central bank's 2026 Financial Stability Report, the information technology sector's share of the S&P 500's total market value is now close to its level at the peak of the dot-com bubble.

That does not mean today's market is the same as 2000.

The major technology companies leading the current AI boom are generally far more profitable and financially stronger than many of the speculative companies that dominated the dot-com era.

But concentration creates its own risk.

When a relatively small number of companies account for a large portion of an index's value, expectations become extremely important.

If those companies continue to deliver exceptional earnings growth, investors may be rewarded.

If growth slows or expected profits fail to materialise, the market can reprice very quickly.

The Bank of Canada has specifically warned that a decline in earnings or earnings expectations among major technology companies could have a significant effect on the broader market.

For Canadian investors who own US equities through ETFs or other investments, this matters even if they do not own individual AI stocks.

The biggest question: who is paying for the AI boom?

One of the most interesting questions surrounding AI is not how much companies are spending.

It is who will ultimately earn enough money from AI to justify that spending.

The world's largest technology companies are investing enormous amounts in data centres, processors, networking infrastructure and energy.

That spending creates revenue for thousands of businesses across the technology supply chain.

But spending by one company is revenue for another company.

It is not automatically new wealth.

Eventually, the AI ecosystem needs to generate enough economic value to justify the enormous capital being invested.

This is where the current investment story becomes more complicated.

AI companies are making impressive claims about future productivity and new business opportunities. But many of the most ambitious applications are still being developed.

The market is therefore attempting to place a value on profits that may not arrive for years.

That can work when expectations are reasonable.

It becomes dangerous when expectations become almost impossible to meet.

The numbers have to catch up with the story

Investors have heard similar stories before.

A new technology arrives.

The potential appears enormous.

Companies invest heavily.

Investors become excited.

Share prices rise.

Eventually, the market begins asking a simple question:

Where are the profits?

That question is becoming increasingly relevant to AI.

Recent reporting has highlighted concerns about the ability of some AI companies to turn rapid revenue growth into sustainable profits.

For example, Reuters recently reported that Anthropic generated $4.6 billion in revenue during 2025 but spent $7.3 billion on compute and infrastructure, while reporting an operating loss of more than $8 billion.

That does not mean Anthropic is destined to fail.

A young technology company can spend heavily today in order to build a much larger business tomorrow.

But it illustrates the scale of the challenge.

AI requires extraordinary amounts of computing power.

Computing power requires data centres.

Data centres require chips, memory, networking equipment, electricity and cooling.

All of that costs money.

At some point, the economic value created by AI applications needs to become large enough to justify the cost of building and operating this infrastructure.

The AI infrastructure race is enormous

The scale of investment is one reason some investors have started using the word "bubble."

According to recent Reuters reporting, global data-centre investment associated with AI could reach extraordinary levels over the coming decades, while economists are questioning whether productivity gains will arrive quickly enough to justify the investment being made today.

That does not mean the investment will be wasted.

In fact, much of today's infrastructure could remain useful for decades.

The railways, electricity networks and internet infrastructure of previous technological revolutions created enormous long-term economic value.

But investors who funded those revolutions did not necessarily all make money.

There is a major difference between:

"This technology will change the world."

and

"This stock is worth its current price."

The first statement can be true while the second is completely wrong.

What happens if AI spending slows?

This is another risk investors should consider.

The current AI ecosystem is interconnected.

Large technology companies spend money building data centres.

Data-centre operators purchase hardware.

Chip companies sell processors.

Memory companies supply the components required by those processors.

Cloud companies provide computing capacity to AI developers.

AI companies use that infrastructure to develop and operate their models.

As long as spending keeps increasing, money flows through the entire ecosystem.

But what happens if the major technology companies decide that they have built enough capacity?

Even a slowdown in the rate of spending could have a significant effect.

It would not necessarily mean AI is failing.

It could simply mean companies are moving from the aggressive investment phase into a period where they want to see stronger returns on the infrastructure they have already built.

That distinction could be important for investors.

A company growing its AI spending by 30% instead of 50% is still spending more money.

But for companies whose valuations depend on extremely rapid growth, that slowdown can matter.

Canadian investors should pay attention

This issue is particularly relevant for Canadian investors because many Canadians gain significant exposure to US technology companies through broad-market investments.

A Canadian investor may not own Nvidia directly.

They may not own Microsoft or Alphabet directly.

But they may own an ETF that holds all of them.

That means investors can have meaningful exposure to the AI trade without thinking of themselves as AI investors.

This is where diversification becomes more complicated.

Owning a broad index is generally very different from owning one technology stock.

But if a growing portion of the index is concentrated in a small group of enormous technology companies, investors should understand what they actually own.

The Bank of Canada's 2026 analysis highlighted precisely this issue, noting that elevated valuations have become increasingly concentrated in a small number of sectors, particularly in the United States.

For Canadian investors, that is a useful reminder to look beneath the label of an investment product and examine its underlying holdings.

This does not mean investors should abandon AI

Calling attention to valuations is not the same as being against artificial intelligence.

Investors should be careful about confusing caution with pessimism.

There is a very strong case that AI will create enormous economic value.

It could make workers more productive.

It could automate repetitive tasks.

It could accelerate scientific research.

It could transform software development.

It could improve healthcare, manufacturing, financial services and countless other industries.

The question is not whether any of this can happen.

The question is how much of that future growth is already reflected in today's stock prices.

That is where disciplined investing becomes important.

An investor does not have to predict whether AI will succeed.

Instead, the investor can ask whether the price being paid provides enough room for disappointment.

The dot-com comparison needs to be used carefully

The easiest comparison for today's AI boom is the dot-com bubble.

But there is a danger in taking the comparison too far.

The technology companies leading the current AI revolution are not comparable in every respect to the speculative internet companies of the late 1990s.

Many of today's leaders have huge revenues, substantial profits, strong balance sheets and established customer bases.

That makes the current environment fundamentally different in several ways.

However, the dot-com era still offers one valuable lesson.

Technological progress does not protect investors from excessive valuations.

Amazon eventually became one of the world's most valuable companies.

The internet was unquestionably transformative.

But investors who bought Amazon at the peak of the dot-com bubble had to endure an enormous decline before the company's long-term success was reflected in the share price.

A great company can be a terrible investment at the wrong price.

That lesson remains relevant today.

Investors should watch earnings, not just headlines

For Martin Becker, this is where investors can separate analysis from excitement.

Rather than simply following the latest AI announcement, investors should watch a few basic measures.

Revenue growth

Are AI-related revenues actually growing?

And more importantly, are they growing quickly enough to justify current expectations?

Profit margins

Revenue growth is not enough.

Companies need to eventually convert that revenue into profits.

If AI increases revenue but also requires enormous increases in infrastructure and operating costs, investors need to understand what happens to margins.

Free cash flow

Cash generation matters.

A company can report strong accounting profits while spending enormous amounts on infrastructure.

Free cash flow provides another way of judging how much money the business is actually generating after investment.

Capital expenditure

Investors should pay attention to how much the major technology companies are spending on AI infrastructure.

The bigger the investment, the bigger the eventual return needs to be.

Valuation

Finally, investors need to ask what expectations are already reflected in the share price.

A company growing earnings at 20% a year can be a fantastic business.

But if the stock price assumes 40% growth, the investment can still disappoint.

The biggest risk may be disappointing expectations

This is perhaps the most important point of all.

Markets do not need AI to fail for AI stocks to fall.

They only need expectations to come down.

Imagine that investors expect an AI company to grow earnings by 40% every year.

The company reports 25% growth.

That is still excellent growth.

But if the market had already priced in 40%, the stock could fall.

This is why market corrections can sometimes appear confusing.

A company can announce strong results and see its shares decline.

The reason is simple:

Investors are comparing reality with expectations, not reality with zero.

The higher expectations become, the harder they are to beat.

What would a healthier AI market look like?

A healthy AI investment cycle would not necessarily mean lower AI spending.

It would mean that spending gradually produces measurable economic returns.

Businesses would use AI to reduce costs or increase revenue.

AI companies would develop sustainable business models.

Infrastructure providers would earn attractive returns without depending entirely on ever-increasing capital expenditure from a handful of customers.

And eventually, productivity gains would become visible across the broader economy.

That would give investors something much more valuable than hype:

evidence.

Evidence that the enormous investment being made today is producing the economic growth investors are expecting tomorrow.

The AI opportunity may survive an AI correction

There is an important irony in the current debate.

Even if an AI bubble develops and eventually bursts, artificial intelligence itself is unlikely to disappear.

The internet did not disappear after the dot-com crash.

Technology did not stop advancing.

Instead, the crash removed some of the excess speculation and allowed investors to distinguish between companies with sustainable business models and those built primarily on expectations.

Something similar could happen with AI.

A major correction would be painful for investors.

But it could also create opportunities.

Companies with strong balance sheets, real customers, growing cash flows and sustainable competitive advantages could emerge stronger.

Meanwhile, companies whose valuations depend almost entirely on future promises could struggle.

Martin Becker's view: focus on the price, not just the promise

For investors, the AI boom presents an unusual challenge.

It is entirely possible to be bullish on artificial intelligence and cautious about AI stocks at the same time.

The technology may be revolutionary.

The economic opportunity may be enormous.

And yet some companies may still be priced too highly.

"The biggest mistake investors can make is thinking they have to choose between being bullish or bearish on AI," Becker says. "The better question is which companies can turn AI spending into sustainable earnings, and whether the current share price already assumes too much success."

That approach may be particularly useful for Canadian investors who have gained substantial exposure to US technology companies through broad-market portfolios.

Rather than trying to predict exactly when an AI bubble might burst, investors can focus on something more practical:

Are earnings keeping up with expectations?

If the answer remains yes, high valuations may be supported.

If earnings begin falling behind expectations, the market could become much less forgiving.

The bottom line

Artificial intelligence could become one of the most important technologies of the 21st century.

There is little reason to doubt that it will change the way businesses operate and how people work.

But investors should remember that markets can get ahead of themselves.

The current AI boom has produced extraordinary investment, extraordinary expectations and extraordinary valuations in parts of the technology sector.

The Bank of Canada has already highlighted elevated valuations, growing market concentration and the potential for AI-related earnings disappointments to trigger a sharp repricing.

That does not mean an AI crash is inevitable.

It means investors should stop asking only:

"How big will AI become?"

They should also ask:

"How much of that future is already priced into the market?"

For investors, that may be the more important question.

Because the difference between participating in a technological revolution and overpaying for it can ultimately come down to one thing:

valuation.

About the Author

Martin Becker is a senior market analyst at NorthernIndex.com, where he focuses on global financial markets, technology and emerging investment trends. His analysis is particularly focused on developments relevant to Canadian market participants, with an emphasis on understanding both the opportunities and risks created by changing market conditions.

Through his work at NorthernIndex.com, Martin examines major economic and market themes to help investors look beyond short-term headlines and make better-informed decisions.