Show Notes
AI CapEx is on track to reach roughly $700 billion at the four largest hyperscalers this year, and equity analyst Irena Petkovic breaks down where all that money is going and what has to be true for it to earn a return. She explains how data centers turned from cost centers into revenue-producing AI factories, the four ways hyperscalers monetize compute, and how the token economy actually works. She then weighs the early evidence of returns against the risks around token pricing, debt financing, and public backlash, and describes how the team positions the portfolio around it.
0:00 - Introduction: The $700 Billion AI CapEx Question 1:29 - How Big Is the Spend? Apollo, Telecom, and Railroads 4:11 - Data Centers as AI Factories: From Cost Center to Revenue 5:53 - Four Ways Hyperscalers Monetize Compute 7:10 - The Token Economy Explained 11:26 - Early Evidence of Returns on AI Investment 14:39 - The Bear Case: Token Prices, Debt, and Backlash 18:49 - Positioning the Portfolio Around AI 20:35 - Outro & Subscribe
Highlights: Hyperscaler AI CapEx of about $700 billion this year is a scale rivaled historically only by the railroads. Capital intensity at Microsoft, Meta, Google, and Amazon has jumped from 5-10% of revenue to upwards of 45%. A high return on invested capital justifies spending down free cash flow rather than protecting it. The data center is now a revenue-producing AI factory that turns electricity and chips into sellable tokens. Compute is monetized four ways: GPU rental, productivity products, enhancing own businesses, and selling tokens. Early returns look encouraging, with sub-three-year hardware payback cited and demand exceeding supply. Risks to watch are token prices falling faster than volumes and a shift toward debt-funded build outs.
Host: Rob Campbell, CFA, Mawer Institutional Portfolio Manager Guest: Irena Petkovic, CFA, Mawer Equity Analyst
This episode is available for download anywhere you get your podcasts.
Founded in 1974, Mawer Investment Management Ltd. (pronounced "more") is a privately owned independent investment firm managing assets for institutional and individual investors. Mawer employs over 250 people in Canada, U.S., and Singapore.
Visit us at: https://www.youtube.com/@MawerInvestment https://www.mawer.com https://www.linkedin.com/company/mawer-investment-management/ https://www.instagram.com/mawerinvestmentmanagement/
More from Art of Boring
Germany's 10-year government bond, negative 0.5% during the pandemic, now yields 3.46%, while the UK's has climbed to 5.5% and Japan's to about 3%. International equity portfolio manager David Ragan examines what the rise in global rates means for how the team values companies. He explains why higher discount rates change the math on future cash flows, how a bank in Ireland became attractively priced once its post-crisis turnaround was overlooked, and why the team treats China's low risk-free rate as artificially depressed. Above all, he returns to a simple idea: the value of diversification.
Q3 2026 market update: Canadian investment-grade bonds returned roughly negative 3% while stock markets in many parts of the world ended September near record highs. Institutional portfolio manager Kevin Minas and investment counsellor Stu Morrow review the quarter, from oil above $100 a barrel and a more hawkish Fed to bond yields at multi-year highs. They discuss why hyperscalers borrowing over US$200 billion this year is crowding out other issuers, what a narrow AI-led rally means when more than half of the S&P 500 sits 20% below its peak, and whether the 60-40 portfolio still needs 40% in traditional bonds. The conversation closes on the quarter's asset allocation changes in the balanced strategy.
Canada-U.S. tariffs are back, and Canadian investors are asking what they mean for their portfolios. Canadian equity portfolio manager Vijay Viswanathan and fixed income portfolio manager Crista Caughlin break down the tariff picture and last month's Canada Investment Summit across stocks and bonds. They cover why the equity portfolio carries little trade-exposed revenue once exempt commodities like oil and potash are set aside, why unemployment matters more than tariffs for Canadian banks, and how a new business-investment tax deduction pulls the effective rate below the U.S. The conversation closes on a simple idea: what the summit needs now is shovels in the ground, not sound bites on TV.
Global credit is absorbing record corporate bond supply as the AI data centre build-out floods the market with debt, yet spreads remain at multi-decade tights. Global credit analysts Sandro Morassutti and Marty Lee revisit the risks the team flagged a year ago, AI-linked issuance, Fed credibility and independence, fiscal spending, and geopolitical risk, and explain why credit spreads have held while government yields have driven the move. They walk through term premium and the long end, why they favour short-dated high-quality credit, the modest duration they have added, and how they are positioned to act when a dislocation finally arrives.
Is AI compute demand durable? Emerging markets equity portfolio manager Wen Quan Cheong returns from a research trip across Taiwan and South Korea and examines AI chip demand, the semiconductor supply chain, and the companies with the strongest structural tailwinds. He breaks compute demand into training, alignment, and inference, explains why margins can widen further down the AI supply chain, and shares what TSMC, SK Hynix, and Samsung revealed about whether this capital cycle is different from the dot-com boom. He also explains how the team trimmed into a crowded AI trade, and makes the case for a non-AI idea: a port operator in the Philippines.