Macro Markets and Machines_The Economic and Market Transformation Driven by AI_GWM report - Flipbook - Page 16
Macro, Markets, and Machines
November 2025
The distinction between the two eras lies in the types of companies we see today. Unlike the dotcom era when
exuberance ran ahead of fundamentals (i.e., earnings), today’s companies, particularly the AI leaders, are
profitable and cash rich, generate consistent free cash flow, and have long histories of earnings growth. AI
investments by the hyperscalers are being largely funded through internal cash rather than through debt
issuance. Their revenue streams are also diversified across cloud services, digital advertising, enterprise
software, hardware, and subscription models, and they possess sustainable competitive advantages relative to
peers, enabling pricing power.
Part of the reason why valuations are stretched today is that the earnings impact of AI investment has yet to
fully materialize. Markets are pricing in future productivity and profitability gains before they have shown up on
financial statements, which is not atypical of early-stage capex cycles. This is not to say that this innovation cycle
will proceed smoothly. Concerns about the sheer scale of the capex cycle are warranted as businesses (and
countries) hurry to get ahead in the AI race. Some capacity may be built out ahead of time, valuations may
overshoot or have already overshot, and corrections/pullbacks are not out of the realm of possibility. But these
factors are unlikely to reverse technological progress, and even though the dotcom bubble burst, it left behind
the technology that powered growth and innovation in the years that followed. The same may apply to AI. Even
if valuations correct, the technology itself is not likely to fade away. It is already being embedded across
industries, supported by governments, and leveraged by consumers and businesses alike.
While much of the valuation expansion has been concentrated among the AI leaders, index-level multiples are
also elevated relative to historical norms, with the S&P 500 currently trading well above its long-term average
forward P/E ratio. Given that valuations are already stretched, further multiple expansion appears unlikely. That
means future market performance will hinge increasingly on earnings growth – and whether productivity gains
can meaningfully lift margins across the corporate sector.
Historically, corporate profit margins and earnings growth have moved in lockstep. As productivity rises and
costs fall, companies can expand margins, which translates into stronger earnings. Historically, for the S&P 500,
a rough rule of thumb is that every 1 percentage point (p.p.) increase in operating margins corresponds with
earnings growth of approximately 7%-8%. This relationship was evident in the dotcom era when rising
productivity led to a >1.5 p.p. rise in operating margins in the five years ending 1999 and annualized earnings
growth of >10%.
Should AI adoption follow a similar path, it could provide meaningful upside to profits, even in a high-valuation
environment. Our work shows that the equity risk premium (ERP) – the additional return investors demand for
holding equities over a risk-free asset, such as Treasuries – stands at ~4% at the time of writing, below the
10-year historical average of ~5%. If the ERP were to rise back to its historical average, the market would be
implying long-term earnings growth slightly north of 6%, just 1 p.p. above the 10-year trailing nominal GDP
growth rate in the United States and below the 10-year historical earnings growth rate. Given that corporate
balance sheets are strong, and companies are generally better capitalized and more efficient than the broader
economy, it seems reasonable to expect earnings growth – particularly for the high-quality players – to handily
outpace nominal GDP growth over time. In other words, equity valuations could be supported if technological
innovation allows firms to lower input costs and expand margins, suggesting markets may be able to “grow into”
today’s valuations.
While valuations and concentration risks warrant consideration, the long-term case for owning equities remains
intact. A general-purpose technology like AI can lift productivity, compress costs and expand profits across a
wide array of industries and sectors. Although the primary beneficiaries are concentrated in a few industries, the
benefits will likely diffuse more broadly as adoption matures, supporting the case for being diversified across
regions and industries to capture the next phase of AI beneficiaries.
Scotia Wealth Management
15
