How Much Influence are Artificial Intelligence Stocks Having Over the S&P 500?

With all the talk of AI stocks, computer chips, and data center construction, many are seeking the answer to the following question: how much is the current market being directly influenced by the proliferation of artificial intelligence?

As of 5/31, we estimate that over 50% of the current market cap of the S&P 500 is comprised of direct beneficiaries of the artificial intelligence and data center investment ramp. This 50% is easy to reach when Nvidia, Microsoft, Alphabet, Amazon, and Broadcom make up approximately 26% of the index. However, the beneficiaries of the AI build out are not limited to technology stocks. Companies like Caterpillar, GE Vernova (leading provider of natural gas turbines and electrical grid equipment), Eaton (electical infrastructure), and Cummins (generators, backup power, electrical systems) have benefitted significantly from the massive ramp in data center construction.

Outside of the top five holdings in the S&P, which collectively made up over 30% of the index as of 5/31, the weight of each subsequent holding quickly drops. These smaller constituents, however, can still have a significant impact when they are all driven by similar economic themes. For example, the S&P currently includes 19 companies from the semiconductors and semiconductor equipment industry. The 3 largest companies, Nvidia, Broadcom, and Micron, make up just under 13% and the remaining 16 combine for another 5%. This 18+% weighting is up from less than 5% as recently as 2020. Since computer chips power the training and execution of these new AI applications, and demand is through the roof, the stocks have soared over the last 3+ years.

Since the end of 2022, we estimate that the top 25 AI beneficiaries (by absolute contribution to the return of the S&P) have contributed around 2/3rds of the total return of the index*. The average annualized return of this group** was 58% vs. 9% for the average stock in the index when excluding the AI beneficiaries. As a whole, the index returned an annualized rate of almost 24% during this time period. Looking at just the largest tech stocks, AKA the Magnificent 7, this small group contributed around 45% of the total return. While the average non-AI stock performed fairly well over the period, boosted by a recovery from 2022's bear market, it is clear that the AI beneficiaries have pulled the overall market return much higher. This surge has not been without warrant given the significant, sometimes massive, growth in revenue and earnings over the period.

With so much of the index being driven by the AI infrastructure build out, many may wonder if the risk of the index has also increased. With over 50% of the index having some direct exposure to one theme, it is hard to say risk has not increased. While there has been some variation in the returns of the largest tech stocks as investors' views of their relative competitive standing in an AI world ebb and flow, recent data shows how much impact shifts in the AI theme can drive markets. According to Bianco Research, the entire gain of the S&P since the end of February has been driven by the "AI Enablers". Additionally, the S&P fell 2.5% on 6/5, but when excluding the AI Enablers basket of stocks, the index was flat on the day. As technology, semiconductors, and other AI beneficiaries eat up a larger and larger weight of the market, the diversification benefit of the non-AI related sectors continues to diminish. The index becomes more dependent on the continued ramp of AI infrastructure investments.

With the massive IPOs of SpaceX, Anthropic, and OpenAI coming down the public listing pipeline, it looks like the market is set to become even more concentrated in AI-related stocks. While SpaceX is conventionally known for its rocket launch and satellite internet businesses, their IPO filing shows that the majority of their stated "total addressable market" (TAM) actually stems from the xAI business. The stated TAM for their AI business is listed at $26.5 Trillion, with $22.7 Trillion coming from enterprise applications. For context, the sum total of revenue generated by all of the companies in the S&P 500 over the last 12 months is currently $19.74 Trillion. Given that the market is so enthralled with AI stocks, it is not too surprising that SpaceX is leaning into the promise of enterprise AI as they look to their public market debut.

The debate continues to rage with plenty of arguments on each side for or against the existence of an AI bubble. Those who argue against the presence of a bubble support it by pointing to the fact that the largest cloud companies (Microsoft, Google, Amazon, and Meta) are funding their massive AI investments out of cash flow rather than out of debt financing or equity capital raises. They contrast this with the lack of earnings and cash flow of many of the dotcom bubble's implosions. In addition, bulls point to the dramatic, and unprecedented, revenue growth rates of Anthropic and OpenAI as evidence of the revolution. Those who argue for the presence of a bubble often point to the price to sales ratios of the AI enablers. As of yesterday, over 30% of the index trades at a valuation that is 10 times their trailing 12 month revenue. This concentration was only around 17% two years ago and the median stock in the index currently trades at just 3.1x sales. Expectations for very high growth and/or very high margins are clearly embedded in the prices of these AI stocks. Lastly, many bubble watchers have noted the multiple circular financing arrangements in which chip suppliers have invested in or financed companies that are ultimately their end customers. They also point to the fact that the semiconductor industry is historically very cyclical but is currently trading like it will grow in perpetuity from the already elevated levels of revenues, earnings, and cash flows. When looking at the facts, the earnings, revenue, and cash flow of the AI beneficiaries certainly paints a different picture than the dotcom era companies being valued how many eyeballs their websites attracted.

The old saying goes that history doesn't repeat, but it does rhyme. I think this saying is as relevant today as ever. While earnings and cash flow are currently supporting the valuations and stock prices of the AI enablers, a shift in optimism or technological innovations that impacts the trajectory of data center build outs would materially impact the market. Such a shift might not result in as drastic a drawdown as experienced in the early 2000s, but that doesn't mean it wouldn't be painful. While the cloud giants have the cash flow to support continued construction, much of this build out appears to be in support of the two largest AI platforms, Anthropic and OpenAI, who are certainly far from supporting their own investments via their cash flow. If there is bubble risk anywhere, it is most likely present where high valuations, cash burn, and lofty projected growth rates exist. This would extend to the venture capital funded AI startups, which attracted over half of VC investments in 2025, that are likely spending a boat load of this fresh capital on AI tools from the large providers.

Commentary from the tech companies in first quarter earnings calls still showed no signs of a slowdown. Companies like Dell noted that customers are racing to order and secure supply given current constraints and rising data center server component prices. This is leading to long lead times and visibility for Dell and peers. Networking company Ciena noted that they are shipping equipment as fast as their customers can put it in the ground and generate data center revenue. In my opinion, Anthropic and OpenAI remain the most important companies to watch in this ever expanding data center build out because (1) they so far need to routinely raise capital to train and deploy their AI models, (2) they are likely the source of a large portion of the incremental growth at the data center owners (Microsoft***, Google, Amazon, Oracle), (3) they are likely generating a substantial amount of revenue from VC-backed AI startups, and (4) any missteps in the sales growth of these companies relative to lofty projections could lead to a ripple effect across any company/industry exposed to the build out of AI data center infrastructure.

Many investors say that this time is different and that a bubble will be avoided in part because of the existing constraints in the supply chain that are preventing excess investment. They may be right, but a bubble avoidance would certainly be an exception to the rule that evolutionary advancements in technology almost inevitably lead to an overbuild of the infrastructure supporting that technology (think railroads, personal computing, and the internet). Given the amount of capital still flowing into AI startups and AI infrastructure investments, it seems the momentum still has plenty of room to run. This does not mean you shouldn't protect against the alternative possibility that enthusiasm is running ahead of reality. We also like to contemplate how an industry like healthcare might be well positioned to capitalize on the efficiencies and advancements of AI despite being left out of the current rally.

It is our core belief that markets are unpredictable and diversification helps protect against overconfidence. As the AI momentum continues to surge, it may be wise to place bets in areas of the market that could be less obvious beneficiaries of the proliferation of artificial intelligence and the resulting boost in efficiency and/or growth. The path of SpaceX shares following the IPO may give good insight into the appetite investors have for continuing to pay eye-watering multiples on the promise of an AI revolution.

*methodology uses starting weight in index as of 12/31/2022 for AI beneficiaries and total return from 12/31/2022 to 5/31/2026 to estimate contribution to market return

**Top 25 AI Contributors includes the following 12/31/22 S&P 500 constituents: AAPL, MSFT, META, NVDA, GOOGL, AMZN, TSLA, AVGO, INTC, MU, LRCX, KLAC, AMAT, CAT, CMI, ETN, TXN, QCOM, WDC, CSCO, AMD, ORCL, CDNS, APH, TT

***Microsoft reported that in its December quarter, 45% of the growth in its remaining performance obligations (contract bookings) was attributable to OpenAI (https://techcrunch.com/2026/01/28/microsoft-earnings-7-6-billion-openai/)