Bank of America Survey Flags AI Capex Boom as Fund Managers Warn of Bubble Risk

Bank of America's latest survey of fund managers has triggered fresh concerns about potential overinvestment in artificial intelligence, with data revealing that AI-related stocks now account for 90% of S&P 500 capital expenditure growth since November 2022. The warning represents a significant shift in sentiment among institutional investors, who are increasingly skeptical about whether the massive spending on AI infrastructure will translate into proportional returns. For the first time in two decades, a plurality of fund managers believe companies are overinvesting in capital expenditures, signaling growing apprehension about the sustainability of the current AI investment cycle.

The implications are stark for investors holding concentrated positions in mega-cap technology stocks. Major artificial intelligence companies including Meta Platforms, Alphabet, Amazon, and Microsoft have underperformed the broader market year-to-date, suggesting that the market may already be repricing some of the risk associated with these elevated capex levels. This divergence between the S&P 500's performance and these AI-heavy stocks underscores the divergence forming within equity markets as investors reassess valuations in light of intensifying capital expenditure demands.

The Capital Expenditure Concentration Risk

The 90% concentration of S&P 500 capex growth in AI-related stocks since November 2022 represents an unprecedented level of capital allocation toward a single technology theme. This degree of concentration is particularly noteworthy given the historical context: fund managers surveyed have not expressed such widespread concern about overinvestment in a single category for the past 20 years. The data suggests that the enthusiasm for artificial intelligence has translated into aggressive capital deployment that may exceed what market fundamentals can justify.

Key metrics from the survey underscore the scale of this phenomenon:

• 90% of S&P 500 capex growth attributed to AI-related stocks since November 2022
• First time in 20 years that fund managers cite overinvestment as a primary concern
• Underperformance of major AI stocks ($META, $GOOGL, $AMZN, $MSFT) versus broader market indices year-to-date
• Significant capital deployment across data centers, semiconductor infrastructure, and AI model development

The concentration risk is particularly acute given the uncertainty surrounding return on invested capital (ROIC) for these massive expenditures. While companies like Microsoft, Amazon, and Alphabet are investing heavily in generative AI infrastructure and capabilities, the timeline for monetizing these investments remains uncertain. Cloud infrastructure buildouts, semiconductor purchases, and data center construction represent multi-year commitments with floating timelines for revenue generation.

Market Context: The Boom-Bust Cycle Concern

The Bank of America survey arrives at a critical juncture for technology valuations. The artificial intelligence sector has commanded an outsized share of market attention and capital flows since the public release of ChatGPT in late 2022, but recent performance data suggests some investors are reassessing their conviction levels. The fact that mega-cap AI stocks have underperformed the market year-to-date despite their dominant capex contributions raises questions about market pricing and expectations management.

Historically, periods of excessive capital investment concentrated in a single theme have often preceded market corrections. The 1990s technology bubble, the mid-2000s housing boom, and the 2017-2018 cryptocurrency surge all shared characteristics of concentrated capital deployment followed by sharp repricing. While no historical parallel is perfect, the 90% capex concentration in AI stocks is reminiscent of those periods where market enthusiasm outpaces fundamental justification.

The competitive dynamics also matter significantly. Major technology companies are engaged in an arms race for AI dominance, with each player compelled to maintain massive capex spending to avoid falling behind competitors. This competitive dynamic can create a self-reinforcing cycle of overinvestment, where companies spend not because returns clearly justify the outlays, but because failing to spend risks competitive obsolescence. The presence of multiple well-capitalized competitors (Microsoft, Google, Amazon, Meta) all pursuing similar AI strategies intensifies this dynamic.

Additionally, regulatory uncertainty surrounding AI governance could impact the ultimate returns on these massive capital investments. Potential regulatory frameworks being debated globally could reshape the economics of AI infrastructure deployment, introducing further uncertainty into already speculative capex decisions.

Investor Implications: Reassessing Portfolio Positioning

For equity investors, the Bank of America survey highlights several critical considerations for portfolio construction. The underperformance of $META, $GOOGL, $AMZN, and $MSFT year-to-date despite their outsized capex contributions suggests that market pricing may be adjusting expectations downward. This repricing could have significant implications for investors with concentrated positions in these names.

The survey's findings suggest a rebalancing strategy that reduces concentration risk in AI-heavy stocks. Strategic allocation toward international equities provides geographic diversification away from the concentrated U.S. mega-cap AI narrative. Markets like Europe and Asia offer exposure to companies with different capex profiles and growth narratives less dependent on artificial intelligence monetization timelines.

Value stock positioning offers another hedge against potential AI bubble risks. Value-oriented businesses—those trading at lower multiples of earnings with established cash flows—are less dependent on speculative future narratives. These stocks tend to outperform during periods of multiple compression, which often accompanies corrections in high-growth technology sectors.

Fixed income allocation serves a critical stabilizing function in portfolios if equity market volatility increases. Bond positions provide portfolio ballast and can fund opportunistic equity purchases if valuations compress sharply. With yields now elevated from the zero-rate environment of recent years, fixed income offers more attractive risk-adjusted returns than it has in over a decade.

The timing of this survey is particularly significant because it captures the perspective of institutional investors who manage substantial capital. When fund managers express concern about overinvestment for the first time in 20 years, that consensus shift often precedes shifts in market behavior. The gap between the concentration of capex growth in AI stocks and their recent market underperformance suggests that repricing may already be underway.

Looking Forward: Managing AI Exposure

The Bank of America survey doesn't argue that artificial intelligence will fail to deliver value—rather, it suggests that the current investment cycle may be front-loading capex with uncertain returns. For investors, this distinction matters greatly. Companies like Microsoft, Google, Amazon, and Meta will likely remain important technology holdings, but their valuation multiples may need to compress to properly reflect return uncertainty.

The prudent approach for most investors involves neither abandoning AI exposure entirely nor maintaining concentrated positions in mega-cap AI stocks. Instead, a diversified portfolio approach that includes international stocks, value equities, fixed income, and selective AI exposure provides resilience against both bubble scenarios and alternative outcomes where AI delivers transformational returns.

As this investment cycle matures, market focus will increasingly shift from capex deployment levels to actual returns on invested capital. Fund managers' concerns about overinvestment serve as a timely reminder that capital discipline matters, even in transformational technology cycles. The sustainability of the current market structure depends heavily on whether the massive AI infrastructure investments generate returns that ultimately justify their costs.