Positioning For AI’s Next Act
The rapid rise of artificial intelligence (or AI) has reshaped market dynamics more profoundly and more quickly than almost any technological shift in recent memory. What began in late 2022 as a breakthrough in accessible AI quickly evolved into a full-scale capital investment race among the world’s largest technology firms. This note explores why AI-related capital expenditures became such a powerful market force, how the funding landscape is now shifting, and what this evolution means for investors. As the AI story enters a more mature—and more selective—phase, understanding who can sustain the race will be essential for navigating the next chapter.
The Story Of Why AI Capex Mattered So Much
For markets, the AI chapter really begins in the autumn of 2022 with the launch of OpenAI’s ChatGPT. From this point, AI stopped being a research story and instead became a mass-adoption reality. While there had been large language models (or LLMs) in service prior to then, ChatGPT’s interface and accessibility meant that anyone could use it as a collaborator. This triggered a capital race as ‘Big Tech’ firms realized that AI represented a clear existential threat to their existing business models.
Several of these firms also understood that the new constraint for AI wasn’t imagination, it was the ability to ‘compute’. After the launch of ChatGPT, usage for it exploded faster than any consumer product in history and the corresponding demand shock meant that technology firms needed to scale up spending to remain in the AI game. However, the cost for training frontier models can be extremely capital intensive. Indeed, only firms with strong balance sheets and cloud scale could afford the additional spending needed to ensure that resources (hardware, software and infrastructure) were adequate for deploying up-to-date AI models.
The uptick in capital expenditures was important for other reasons. For one, it meant a scramble to secure supply of GPUs and to overbuild cloud infrastructure deliberately. This was done to ensure as little of a dependence as possible on other hyper-scalers, but in some instances, this couldn’t be avoided. Consider how well the share prices of firms that had already specialized in GPU manufacturing (Nvidia, AMD) have done over the past few years.
Second, the degree of spending in the AI space became an economic story. The most direct way to proxy the effect of AI on the economy is to look at the growth in spending on software and research/development in the GDP numbers that the Bureau of Economic Analysis (BEA) releases. The amount of spending in those categories has averaged 7.8% quarterly growth (on an annualized basis, we might add) over the past year – which is roughly three times the pace that headline real GDP has grown at over that same timeframe.
Third, and most importantly for markets, there was a keen sense that any investment in AI should be rewarded. Part of the reason for this was that spending on AI was taken to mean that tech firms were serious about AI and would do whatever was necessary to keep up in the race. This was an ideal passive environment as AI-related capital expenditures were taken as a signal and not just an investment.
For Markets – Funding Is The Key
For any tech firm investing in the space, the dirty secret about AI is there is no reliable way to quantify the value-add. Of course, we know that AI is a force multiplier for productivity across several sectors, but making the connection to the amount being spent on building out the infrastructure is a bit trickier to do. This is important, because as the degree of spending in this space ramps up, questions about how firms will fund this spending will continue to linger.
In the period between the autumn of 2022 and early 2025, there wasn’t as much of a concern about this. After all, the mega-cap tech firms in the AI race could fund the infrastructure spend via operating cash flows. This is a crucial detail that distinguishes this current AI frenzy from the “tech bubble” 25 years ago. Tech firms back then couldn’t rely on endogenous cash flows and instead had to tap into external financing (IPOs mainly) to fund themselves. By contrast, the tech firms in the current AI race had fortress balance sheets and were flush with cash.
However, a problem arises when the pace of spending is increasing relative to the growth in cash flow from operations. Once this happens, then firms invested in the AI race cannot rely on endogenous cash flows alone. Instead, it has to look at other forms of financing.
Enter debt issuance. Since the start of the second half of 2025, the ‘hyper-scalers’ have been tapping debt markets as a source of funding at a far more aggressive pace than in prior years. Why debt over equity financing? Because debt is cheaper to issue, avoids the stigmas around equity issuance, and hyper-scaler balance sheets can easily accommodate them.
However, reliance on ever‑larger capex budgets creates an asymmetry of risk for equity holders. If a meaningful share of AI investments proves non‑economic, assets will be written down while the nominal value of newly issued debt remains fixed—tightening financial flexibility and amplifying downside to equity. As spending growth outpaces operating cash flow growth, the margin for error narrows. This is why investors have shifted from passively rewarding AI spend to actively demanding capex discipline: management teams must now demonstrate hurdle‑rate rigor1, clearer paths to monetization, and pacing that aligns with measurable utilization rather than aspirational scale.
Where Does The AI Story Go From Here
Given the above, markets have become even more sensitive to planned expenditures going forward. The lofty valuations for the ‘Mag 7’ tells us that there is a fair bit of optimism already in the price. The narrative has shifted away from rewarding big spending to prioritizing clarity and resilience amidst potential disruption.
More directly, what this means is that AI investing is shifting from a passive ‘beta’ story to a more active one.2 Said differently, there will now be winners and losers in the space going forward. Determining who wins and who loses will likely come down to a few factors including:
- The degree by which firms are reliant on external debt financing to fund big projects.
- Understanding which firms are more adaptive and resilient to AI advances.
Going forward, we expect the hierarchy in the AI investment race to be restructured. At the top will be firms that have capital structures with the most capacity to absorb debt, and with less exposure to potential advances made in the AI space. The latter point is important— consider the performance of firms that have been heavily reliant on software as part of their business models of late. We do expect that markets will treat additional spending by such firms more judiciously than others.
Tools To Help You Stay Invested
Given this pivot from “reward spending” to “reward discipline,” the opportunity set is broadening beyond a handful of hyperscalers to companies that can apply AI productively and finance growth prudently. That shift argues for portfolio constructions that capture leadership diffusion while managing single‑name concentration and currency risk.
First, we expect leadership to continue broadening as AI adoption lifts productivity in non‑tech sectors—industrial automation, healthcare workflows, and materials logistics among them. As earnings dispersion narrows and valuation gaps compress, equal‑weight exposures can better capture this diffusion than cap‑weighted indices that remain top‑heavy.
Second, for investors seeking to express targeted views on potential AI “winners” while managing FX, BMO Global Asset Management’s Canadian Depositary Receipts (CDRs) provide CAD‑hedged access to select U.S. names, including several hyper-scalers.
Third, for investors that are bearish on the AI space as a whole, there are some sectors that provide ‘hedge-like’ behaviour during periods when tech and telecoms are on the defensive. Since the end of September, we’ve noted that the healthcare, materials and industrials have done particularly well (for different reasons). Each of these sectors are likely to incorporate aspects of AI to improve productivity going forward.
As AI investment enters a new phase, markets are shifting from rewarding sheer spending to scrutinizing the sustainability and strategic value of that spending. The next stage of AI leadership will be defined not only by innovation, but by balance sheet strength, adaptability, and the ability to withstand competitive disruption. For investors, this transition presents opportunities across sectors as AI adoption broadens and market leadership diversifies. Whether through equal weight strategies, targeted CDR exposure, or selective sector hedges, positioning thoughtfully will be essential. The AI race is far from over—but its winners will be chosen by discipline as much as ambition.
Bipan Rai joined BMO Global Asset Management in 2024 and currently serves as Head of ETF and Alternatives Strategy. He is highly regarded for his macroeconomic insights as well as his knowledge of market structure for various asset classes. Mr. Rai holds an MBA from the Schulich School of Business at York University and a Bachelor of Engineering degree (Aerospace Engineering) from Toronto Metropolitan University.
1 Hurdle-rate rigor refers to the disciplined application of a minimum required rate of return (the "hurdle rate") to filter, prioritize, and approve investment projects or capital expenditures. It acts as a financial "litmus test," ensuring that a firm only commits capital to opportunities that are projected to exceed its cost of capital and adequately compensate for specific risks.
2 Beta is a measure of a security's or portfolio's volatility or systematic risk in comparison to the market as a whole. It tells you how much a security’s price is expected to swing relative to broader market movements.
An investment in Canadian depositary receipts (“CDRs”) may not be suitable for all investors. Each series of CDRs is only offered to investors in Canada in accordance with applicable laws and regulatory requirements.
Important information about these investments is contained in the short form base shelf prospectus and prospectus supplement for each series of CDRs (together, the “Prospectus”). Purchasers are directed to www.sedarplus.ca or to bmogam.com to obtain copies of the Prospectus and related disclosure before purchasing CDRs. Each series of CDRs relates to a single class of equity securities (the “Underlying Shares”) of an issuer incorporated outside of Canada (the “Underlying Issuer”). For each series of CDRs, the Prospectus will provide additional information regarding such series, including information regarding the Underlying Issuer and Underlying Shares for such series. Neither BMO and its affiliates nor any other person involved in the distribution of CDRs accepts any responsibility for any disclosure provided by any Underlying Issuer (including Information contained herein or in the Prospectus that has been extracted from any Underlying Issuer’s publicly disseminated disclosure).
Disclaimer:
The BMO ETFs sponsored articles are intended for information purposes only. They have been prepared by the respective authors and represent their assessment at the time of publication. The comments herein do not necessarily represent the views of BMO Global Asset Management. The views are subject to change without notice as markets change over time. The information contained herein does not constitute a solicitation of an offer to buy, or an offer to sell securities, and should not be construed as investment, tax or legal advice to any party. Particular investments and/or trading strategies should be evaluated and professional advice should be obtained with respect to any circumstance. Past performance is no guarantee of future results.
Any statement that necessarily depends on future events may be a forward-looking statement. Forward-looking statements are not guarantees of performance. They involve risks, uncertainties and assumptions. Although such statements are based on assumptions that are believed to be reasonable, there can be no assurance that actual results will not differ materially from expectations. Investors are cautioned not to rely unduly on any forward-looking statements. In connection with any forward-looking statements, investors should carefully consider the areas of risk described in the most recent simplified prospectus.
Commissions, management fees and expenses all may be associated with investments in exchange-traded funds. Please read the ETF Facts or prospectus of the BMO ETFs before investing. Exchange-traded funds are not guaranteed, their values change frequently and past performance may not be repeated. For a summary of the risks of an investment in the BMO ETFs, please see the specific risks set out in the prospectus. BMO ETFs trade like stocks, fluctuate in market value and may trade at a discount to their net asset value, which may increase the risk of loss. Distributions are not guaranteed and are subject to change and/or elimination.
"The Select Sector SPDRÆ Trust consists of eleven separate investment portfolios (each a “Select Sector SPDR® ETF” or an “ETF” and collectively the “Select Sector SPDR® ETFs” or the “ETFs”). Each Select Sector SPDRÆ ETF is an “index fund” that invests in a particular sector or group of industries represented by a specified Select Sector Index. The companies included in each Select Sector Index are selected on the basis of general industry classification from a universe of companies defined by the S&P 500Æ. The investment objective of each ETF is to provide investment results that, before expenses, correspond generally to the price and yield performance of publicly traded equity securities of companies in a particular sector or group of industries, as represented by a specified market sector index.
The S&P 500®, SPDRs®, and Select Sector SPDRs® are trademarks of The McGraw-Hill Companies, Inc. and have been licensed for use. The stocks included in each Select Sector Index were selected by the compilation agent. Their composition and weighting can be expected to differ to that in any similar indexes that are published by S&P.
The S&P 500 Index is an unmanaged index of 500 common stocks that is generally considered representative of the U.S. stock market. The index is heavily weighted toward stocks with large market capitalizations and represents approximately two-thirds of the total market value of all domestic common stocks. The S&P 500 Index figures do not reflect any fees, expenses or taxes. An investor should consider investment objectives, risks, fees and expenses before investing.
You cannot invest directly in an index.
An investment in Canadian depositary receipts (“CDRs”) issued by Bank of Montreal (“BMO”) may not be suitable for all investors. Important information about these investments is contained in the short form base shelf prospectus and prospectus supplement for each series of CDRs (together, the “Prospectus”). Purchasers are directed to www.sedarplus.ca or to bmogam.com to obtain copies of the Prospectus and related disclosure before purchasing CDRs. Each series of CDRs relates to a single class of equity securities (the “Underlying Shares”) of an issuer incorporated outside of Canada (the “Underlying Issuer”). For each series of CDRs, the Prospectus will provide additional information regarding such series, including information regarding the Underlying Issuer and Underlying Shares for such series. Neither BMO and its affiliates nor any other person involved in the distribution of CDRs accepts any responsibility for any disclosure provided by any Underlying Issuer (including information contained herein or in the Prospectus that has been extracted from any Underlying Issuer’s publicly disseminated disclosure). Each series of CDRs is only offered to investors in Canada in accordance with applicable laws and regulatory requirements.
Commissions, management fees and expenses all may be associated with investments in exchange-traded funds. Please read the ETF Facts or prospectus of the BMO ETFs before investing. The indicated rates of return are the historical annual compounded total returns including changes in unit value and reinvestment of all dividends or distributions and do not take into account sales, redemption, distribution or optional charges or income taxes payable by any unitholder that would have reduced returns. Exchange-traded funds are not guaranteed, their values change frequently and past performance may not be repeated.
BMO ETFs are managed by BMO Asset Management Inc., an investment fund manager, a portfolio manager, and a separate legal entity from Bank of Montreal. BMO Global Asset Management is a brand name under which BMO Asset Management Inc. and BMO Investments Inc. operate.
BMO (M bar roundel symbol) is a registered trademark of Bank of Montreal, used under licence.