Given the strength of market gains over the last three years and the ambiguity surrounding the return potential from the sizable spending associated with AI infrastructure, we believe a reset in equity expectations toward a more moderate return expectation is appropriate.
A surge in stock prices of over 30% since the April lows1, along with growing headlines about sizable AI infrastructure investments, has stoked concerns about whether a rising bubble is developing in the equity market. This is a natural query given the circumstances, but the answer is not straightforward, as pockets of excess are conditional on the ultimate evolution of AI. In this assessment, we aim to address stock valuations and how we view the scenarios for AI build-out influencing the equity market.
Despite the economic uncertainty associated with tariffs and other geopolitical shifts in 2025, equity prices have continued to trend higher, with the S&P 500 up +14% on the year2. If you look back further, to the market lows in October 2022, the S&P 500 has generated a stellar 24% annualized return3, a level nearly three times its long-term average. This robust appreciation has stemmed from a combination of earnings growth in the underlying companies in the S&P 500, as well as higher valuations, as investors are willing to pay more for future growth, particularly among companies deemed likely beneficiaries of AI.
Where that leaves us today is that valuations sit at above-average levels for large companies. If you isolate this more specifically into higher-growth companies, as denoted by the Russell 1000 Growth index, valuation multiples are near where they peaked in 2021, but still below the levels experienced in 1999, before the technology bubble. Given this, it’s not a stretch to draw a corollary and conclude there is increased risk in equity markets. This may be fair, but there is much more to the story than the valuation of the broad index.
The high valuation of these broad indexes stems from the elevated valuations of leading technology companies, notably the Magnificent 7 stocks (Apple, Amazon, Alphabet, Meta, Microsoft, NVIDIA, Tesla). These stocks trade, on average, at a price-to-earnings (P/E ratio) multiple of 36x earnings4. This is undoubtedly an elevated valuation, considering the S&P 500’s long-term average price-to-earnings multiple is around 17x. However, one should consider that the earnings of these companies increased 27% over the past year5, compared to the S&P 500, which, over the long term, averages 7% growth. A rudimentary way to account for growth rates when considering valuation is to divide the P/E ratio by the growth rate6. For Mag 7 companies, this ratio is 1.3x, which currently stands at a reasonable level. This perspective, on the other hand, is predicated on the notion that the current growth rate does not begin to slow materially.
The risk of a slowing growth rate for AI-fueled stocks is a real consideration. However, that risk seems less likely in the near term, and we believe it is more likely to be company-dependent rather than industry-wide. The number of companies competing to develop AI models capable of learning and inference is substantial. It not only includes models being developed by large, publicly traded companies but also newer, non-public entrants, such as OpenAI, Anthropic, xAI, and others. While it’s uncertain how the AI build ultimately plays out, our expectation is that not all companies will share in material economic benefits from AI. Following other disruptive innovations from the past, such as automobiles or search engines, the number of competitors is likely to narrow over time, and not every company will reap a fair share of the profits.
The narrowing of competition is what we expect will lead to areas of excess in AI. At this point, technology companies continue to build, train, and refine AI models as inference use cases are only beginning to emerge. AI model developers recognize they need substantial infrastructure and compute power to adequately scale their business should customer demand for their AI solutions accelerate. The need to have this future capacity in place is what is fueling the recent surge in planned data center investment. In fact, Meta CEO Mark Zuckerberg commented this fall that he’d rather misspend a couple of hundred billion dollars than lose the AI race. Given Meta’s balance sheet, they can afford this risk.
The hazard, as we see it, is that companies that fail to achieve the anticipated level of AI model adoption will be left with excess data center capacity. If we are accurate in our assessment that not all AI models will achieve commercial success, then at some point (many years into the future), we will likely face an environment with excess AI infrastructure. In addition, those companies that do not achieve success will have to grapple with the fact that they made investments, potentially in the hundreds of billions of dollars range, that will earn little to no return. The return on investment from those successful AI companies is not yet clear, but we expect the returns to be reasonable.
The challenge as we sit here today is to predict who the ultimate winners will be, which can’t be done with certainty. Therefore, our approach at this point in the AI development cycle is to maintain broad exposure to large tech companies and more proactively focus portfolio exposures in areas of the AI buildout that are less binary, such as those related to power demand. Alternatively, this perspective has led us to mitigate our exposure to data centers, which we view as having greater property-specific risk if the industry ultimately ends up with excess capacity from AI players that don’t achieve commercial success.
Much of the discussion and debate surrounding market excess centers on AI and the associated technology companies investing in this technology. If we look beyond these companies and at the broader market, we find valuations are more reasonable and more closely aligned with long-term averages. The divergence in valuations and growth rates serves as a reminder of the importance of portfolio diversification. Reflecting on the late 1990s, when there was an excess of internet stocks and infrastructure, the pullback in equity prices did not impact all stocks equally. Investors with exposure to higher-quality businesses or more attractively valued stocks were able to garner returns superior to those of broad-based U.S. large-cap equities, with a notable dispersion of stock returns. The lesson learned from this period was that diversification mattered, and not leaning entirely into an attractive trend served as the best strategy over the long term.
Conclusion
Given the strength of market gains over the last three years and the ambiguity surrounding the return potential from the sizable spending associated with AI infrastructure, we believe a reset in equity expectations toward a more moderate return expectation is appropriate. The opportunities that will stem from AI and other long-term thematic trends provide a solid basis for growth and investment opportunities. Not all companies will benefit, and we are entering a market environment where security selection and consideration around risk management will play a greater role than they may have in recent years. We are optimistic about the opportunities ahead, but we remain pragmatic that trends and market conditions will continue to evolve, and we will continue to evaluate the right mix of assets in our clients’ portfolios.
Sources:
- Based on the return of the S&P 500 index from April 8, 2025, through November 19, 2025.
- As of November 19, 2025.
- Source: Bloomberg, 10/14/2022 through 11/19/2025.
- Based on the Bloomberg Magnificent 7 Index, as of November 19, 2025.
- Source: BlackRock. As of November 17, 2025. https://www.blackrock.com/us/individual/literature/market-commentary/weekly-investment-commentary-en-us-20251117-at-last-key-us-economic-data-return.pdf
- Referred to as a P/E to growth rate or PEG ratio
Disclosures: Composition Wealth LLC (“Composition”) is a registered investment advisor. Advisory services are only offered to clients or prospective clients where Composition and its representatives are properly licensed or exempt from licensure. The views expressed in this commentary are subject to change based on market and other conditions. These documents may contain certain statements that may be deemed forward‐looking statements. Please note that any such statements are not guarantees of any future performance and actual results or developments may differ materially from those projected. Any projections, market outlooks, or estimates are based upon certain assumptions and should not be construed as indicative of actual events that will occur. The information is illustrative, provided is for educational and informational purposes only and does not constitute investment advice and it should not be relied on as such. It should not be considered a solicitation to buy or an offer to sell a security. It does not take into account any investor’s particular investment objectives, strategies, tax status or investment horizon. You should consult your attorney or tax advisor. All information has been obtained from sources believed to be reliable, but its accuracy is not guaranteed. There is no representation or warranty as to the current accuracy, reliability or completeness of, nor liability for, decisions based on such information and it should not be relied on as such. Asset Allocation may be used in an effort to manage risk and enhance returns. It does not, however, guarantee a profit or protect against loss. No investment strategy or risk management technique can guarantee returns or eliminate risk in any market environment. All investments include a risk of loss that clients should be prepared to bear. The principal risks of Composition’s strategies are disclosed in the publicly available Form ADV Part 2A. Past performance shown is not indicative of future results, which could differ substantially.