Rocket ships to the moon

With contributions from Noha Fazili, research analyst

Please note: Unless otherwise noted, all content and data are as of June 25, 2026.

Markets are over the moon as AI hopes overcome geopolitical fears

In a continued show of resilience, equity markets have pushed higher on a year-to-date basis, particularly in North America. In fairness, U.S. equity markets did correct roughly eight percent with the onset of the U.S. and Israeli strike against Iran, and bottomed on March 30. Since then, however, markets have staged what looks like a V-shaped recovery, bouncing back more than 15 percent and soaring back above preconflict levels to reach new highs.

The euphoric debut of SpaceX, the highly anticipated IPOs from Anthropic and OpenAI, alongside the Memorandum of Understanding (MoU) between Iran and the U.S., have buoyed the markets’ spirits.

Corporate earnings engine keeps chugging along smoothly

The equity markets’ push higher can be credited to strong corporate earnings, the AI investment wave and an economy that has thus far absorbed higher energy prices better than expected. The one caveat to this is that the rebound from the March lows has not been broad-based, with Energy, Technology and Communication Services sector stocks leading the charge in an otherwise relatively narrow market.

S&P 500 sector returns

The hardiness of the corporate earnings growth engine was evident in the first quarter of this year. Q1 earnings reports for the S&P 500 index have concluded, with more than 85 percent of companies beating consensus estimates—the highest beat rate since Q2 2021 , according to FactSet. Earnings growth during the quarter notched higher by roughly 28 percent—the highest since Q4 2021—and the sixth consecutive quarter of double-digit earnings growth for the S&P 500 Index , according to FactSet. Not surprisingly, the Information Technology sector posted the highest earnings growth of all 11 sectors in the S&P 500 at more than 50 percent, powered predominantly by the (seemingly) indomitable AI trend.

Earnings growth expectations for 2026 have been revised higher from the low-to-mid teens at the beginning of this year to north of 20 percent. Consensus expectations are for another 10-to-15-percent growth in earnings in 2027. As a recurring theme—and something of a hallmark of a highly concentrated index—the positive earnings revisions have been quite concentrated in Information Technology, and this sector is forecast to lead the charge and outpace the broader market from an earnings growth perspective over the next two years.  

U.S. equity market valuation in context

Valuations for the S&P 500 index currently sit above its historical average and above developed-market peers. But some context is perhaps warranted. First, U.S. equity market valuations have been disproportionately driven by earnings growth rather than multiple expansion. To wit, the current price-to-earnings (P/E) ratio of roughly 21x sits below the recent peak of 23x . Second, excluding the Magnificent Seven (Microsoft, Amazon, Alphabet, Meta, Nvidia, Tesla and Apple) from the index’s valuation—to control for the close to 30 percent concentration of this block of stocks in the S&P 500 index—yields a valuation more or less in line with the long-term average.

The valuation of the S&P 500 is much more reasonable without the Magnificent Seven

Lastly, the higher valuation for the U.S. equity markets relative to peers is, we believe, in part justified by the higher-than-expected earnings growth for the market on a relative basis versus other developed markets in 2026 (with the exception of Canada) and 2027.

We believe it will be hard for investors to bet against the U.S. equity markets given they lead the charge on the AI revolution. A modest overweight in this asset class is reasonable in our view, though investors should be wary of the noted concentration in the index. Portfolio discipline around managing valuation risk, rebalancing and diversification (across equities and asset classes) remains critical.

Code and capital: Assessing the AI arms race

One of the key underpinnings of the growing concentration in technology in U.S. equity markets has been the AI arms race—dominated by a handful of mega-cap hyperscalers. Capital expenditure (capex) across the top hyperscalers has been nothing short of historic, with the sector on track to surpass an estimated $650 billion in spending this year . This is 80 percent higher than 2025 levels and marks a continuation of one of the fastest and largest capex cycles in decades.

Hyperscalers are expected to continue ramping up capex

The level of capex spending announced for 2026 currently represents two percent of U.S. GDP, with increases likely in the next two years—consensus AI capital expenditure estimates suggest this could rise above $1 trillion in 2027 . The unprecedented level of spending, coupled with the relatively high valuations for AI-related and adjacent stocks, has evoked comparisons with the internet bubble of the late ‘90s. Yet we view this as the first inning of an AI supercycle, rather than a bubble. For one thing, it is being funded predominantly by corporate earnings rather than speculative debt.

The funding is concentrated disproportionately among established “Big Tech” mega-cap companies with strong cash-flow generation and robust balance sheets—not cash-burning start-ups. Through the lens of capex margins, even with the groundbreaking levels of spending, these remain below where they were during the Dot-Com bubble and the Gilded Age.

Capex as a percent of revenue

One of the other attributes of this technological revolution is scarcity—not excess—related to a structural shortage in computing power, electrical-grid capacity, data centres and high-bandwidth memory. Put differently, a massive buildout of physical infrastructure to support the development of AI technology and high demand is in its early stages, with few signs of overbuild. Rather than spending on abstract ideas, the outlays of cash are an investment in physical infrastructure intended to support growing demand .

Accelerating demand for data centres*

U.S. new leasing (MW)

There is no question that investors’ focus going forward will be on the return on investment for these gargantuan capex outlays. Positively, in recent years large language models (LLMs) have continued to demonstrate improvements of more than sixfold per year , according to RBC Global Asset Management.

AI models are becoming exponentially more capable

RBC GAM also estimates that AI infrastructure expenditure contributed roughly 0.5 percent to U.S. GDP growth in 2025 —an important partial offset to tariff-related economic growth headwinds of -0.75 percent. A similar contribution from AI growth is pencilled in for 2026.

Enterprise adoption of AI has been on the rise worldwide. According to a report by McKinsey, AI adoption will reach just under 20 percent among the working-age population by the end of 2026 . In the corporate sector, more than 85 percent of companies are using AI in at least one business function, although much work remains in the pilot or experimentation phase. Over 60 percent of the businesses surveyed in the report expect AI to directly enhance productivity levels, albeit a third expect workforce disruptions in certain roles owing to greater AI-driven automation. RBC Global Asset Management points to an acceleration in U.S. productivity growth since the introduction of ChatGPT, suggesting that the benefits of this transformational technology are already being felt.

U.S. labour productivity growth has accelerated since ChatGPT’s release

With respect to future monetization potential, Wall Street analysts point to the estimated $2 trillion backlog across leading hyperscalers . This backlog represents future contracted cloud revenue driven mainly by accelerating enterprise AI infrastructure demand. The real test will be whether the high levels of expenditure drive computational efficiency that ultimately leads to sizable margin expansion for these companies. In the short-to-medium term, we believe it is fair to expect some margin compression before meaningful margin expansion unfolds over the long term.

Ultimately, the promise of AI to reshape global productivity is compelling and echoes the transformative waves of the steam engine and the internet. Yet, history has shown that the paths of progress born of such technological revolutions are riddled with fits and starts, and do not unfold in a linear fashion. Just as the rollout of electricity or early broadband internet encountered massive infrastructure hurdles, market volatility and some painful shakeouts before fundamentally altering our economy, the AI revolution is likely to face its own periods of turbulence.

Investors must navigate the inevitable volatility with discipline. The most enduring portfolios will be built by those who balance exposure to AI’s exponential potential with the pragmatic understanding that the path to widespread adoption is always paved with growing pains.

---

---