In an era of AI-driven enthusiasm and fear, maintaining humility and positioning portfolios with differentiated drivers and resilient themes may be the best defence against volatility.
Artificial intelligence remains a dominant market force, supported by continuous innovation and rising adoption by enterprises and consumers. Yet the debate has now shifted from whether AI will transform the economy (it has) to where the economic value created by that transformation will ultimately accrue.
The initial phase of the AI buildout, which often rewarded mere exposure to the theme, is giving way to a mature phase in which investors demand evidence that AI spending can translate into durable earnings and cash flows.
This distinction carries important portfolio implications given AI’s growing influence across markets. Greater scrutiny of company execution now carries broader consequences in indices dominated by a handful of AI leaders, magnifying the impact of a pullback during periods of softening sentiment. As AI’s footprint expands across sectors and regions, navigating the next phase of the cycle will likely require greater discipline and a deeper understanding of where value is being created and where risks are accumulating.
Three shifts are reshaping the AI investment landscape: the market is placing greater emphasis on execution, AI-related risks are becoming more influential in market outcomes, and AI exposure is spreading well beyond the technology sector. Together, these developments are increasing the value of selectivity.
1. The market is moving from AI promise to AI delivery
The shift in investor focus from broad exposure to execution is already reshaping the dynamics of the tech sector. As the AI narrative evolves from promise and potential to implementation and delivery, expectations for the tech sector are becoming more demanding. This comes at a time when forecasts remain ambitious, with the tech sector expected to deliver roughly 60% earnings growth this year. Yet those expectations are colliding with growing concerns around the scale of hyperscaler investment, intensifying cash flow pressures, and a less supportive cyclical backdrop. Together, these dynamics leave less room for disappointment.
Such an environment helps explain the rising performance dispersion within AI’s leading companies. A landscape that once rewarded mere participation in the AI buildout has become less forgiving, with investors less willing to reward AI exposure without clear evidence of monetisation, margin expansion, and durable cash-flow generation.
The implications of this shift in investor sentiment extend beyond individual company performance. Rising scrutiny toward tech would be less consequential if AI exposure were broadly distributed across the market. Instead, equity market performance has become more dependent on a relatively small group of AI-linked companies, amplifying the impact of both positive and negative AI-related developments.
2. AI concentration is amplifying market risks
With roughly 40% of the S&P 500 concentrated in just ten AI-centric stocks, the cap-weighted index has become highly vulnerable to developments affecting its market leaders, from idiosyncratic risks to fluctuations in AI sentiment. AI’s drag on the index during this summer’s tech selloff, along with other sharp pullbacks since AI-bubble concerns emerged last fall, underscores the pitfalls of elevated market concentration.
What’s more, this dynamic is not limited to U.S. equity markets. Interconnected global tech supply chains have also amplified AI’s influence across global equity benchmarks.
Three leading Asian semiconductor companies, which are critical suppliers to U.S. hyperscalers, now account for nearly a third of the MSCI Emerging Markets index. Their outsized weight in the cap-weighted index is reducing the diversification benefits traditionally associated with emerging markets while heightening global equities’ vulnerability to periods of AI-related pullbacks.
Even more challenging for investors from a diversification perspective, AI exposure is no longer confined to big tech. AI’s complex supply chains are creating deeper dependencies between tech and a broad range of industries and regions often viewed as diversifiers, leaving portfolios more exposed to AI-related volatility.
3. AI exposure extends far beyond technology
Growing demand for data centres and supporting AI infrastructure has boosted demand for power, electrical equipment, cooling systems, and commodities, creating new beneficiaries across industrials, utilities, materials, and energy sectors globally. As a result, regions often viewed as less exposed to AI, including Europe and Latin America, now have greater exposure to the structural theme through infrastructure providers and commodity producers that support the global AI buildout.
As AI exposure spreads across sectors and regions through interconnected supply chains, portfolios that appear diversified may be more exposed to AI than investors realise. During the early stages of the cycle, that exposure was advantageous; screening for AI exposure was often enough to determine which companies benefited from the buildout. However, as AI’s influence becomes more embedded in the global economy and AI-related volatility becomes more frequent, building resilient portfolios is both more important and more challenging, requiring greater discipline from investors.
Therefore, while sectoral and regional diversification remains important, investors also need to look deeper at the underlying drivers of earnings, particularly where AI-related exposure is less obvious. Specifically, a greater understanding of which companies can translate AI exposure into durable earnings growth, and which may struggle to convert AI-driven investment into lasting value creation, will be essential for identifying the next phase of winners and losers.
From broad exposure to company selection
As investors place greater emphasis on execution and tangible results, markets are becoming increasingly focused on distinguishing AI enthusiasm from fundamental strength. While stock pickers have faced headwinds during a period when market returns have been largely driven by a narrow group of AI beneficiaries, a more demanding phase in the AI cycle, marked by rising performance dispersion, may place greater value on rigorous company-level analysis in three key areas:
Separating temporary from durable AI beneficiaries
Not all AI infrastructure stocks are positioned to capture durable earnings growth created by the AI buildout. While some companies may currently benefit from supplying one-off data centre components amid the spending boom, others may be better positioned to capture sustained demand if and when the spending slows. While hyperscalers have not indicated any shift in spending guidance, against this uncertainty, stock selection based on earnings durability, rather than trying to predict the timing of any inflection point, can be prudent.
Identifying adoption winners
Company-level analysis may also help identify firms positioned to benefit from AI adoption, particularly as investors shift attention from AI enablers and infrastructure stocks to companies realising meaningful productivity gains and margin improvement through AI adoption. As executives and leadership teams navigate how to incorporate AI into their businesses, productivity gains will likely vary from company to company, making them difficult to identify through index exposure. This makes company execution a key consideration requiring active judgment, particularly as sectoral hurdles—such as strict regulatory requirements in healthcare and financials—could widen performance differentials, rather than AI acting as a rising tide that lifts all boats.1
Capitalising on current market dislocations
Broad and indiscriminate selloffs across segments perceived as vulnerable to AI disruption, namely software and services, may create opportunities where expectations have become overly pessimistic. Investors willing to conduct rigorous analysis may uncover whether a company can retain its competitive moat, adapt its business model, and scale alongside industry innovation, or fail to adapt and become obsolete. Distinguishing the likely winners and losers in an oversold segment is an area where active management can still add value.
Implications for investors
Elevated market concentration and expanding AI linkages across industries and regions require greater discipline in building resilient portfolios. As AI influences a growing share of corporate earnings across sectors and geographies, diversification may increasingly depend on underlying business drivers rather than traditional regional or sector allocations.
While the development phase of the AI cycle rewarded exposure, the maturing phase is likely to reward selectivity. In that environment, stock-picking strategies will be increasingly important in identifying durable beneficiaries, adoption winners, and businesses whose long-term prospects are not fully reflected in current valuations.
In an era of AI-driven enthusiasm and fear, maintaining humility and positioning portfolios with differentiated drivers and resilient themes may be the best defence against volatility.
By Magdalena Ocampo, Market Strategist at Principal Asset Management
