Key Takeaways:
The traditional 60/40 stock-bond portfolio is obsolete because artificial intelligence has become impossible for investors to avoid, according to Apollo's chief economist.
For decades, the golden rule of investing was simple: Put 60% of money into stocks and 40% into bonds. That framework is officially dead, replaced by a binary bet on artificial intelligence versus everything else, according to Torsten Slok, chief economist at Apollo.
"The new 60-40 is AI vs. non-AI," Slok said in a research note shared with MarketWatch.
The 10 largest companies in the S&P 500 now account for about 40% of the capitalization-weighted index's value, Slok said. Nine of those 10 have businesses tied to AI, with drugmaker Eli Lilly the lone exception. Nvidia Corp. alone represents 7.5% of the benchmark, followed by Apple Inc. at 6.8% and Alphabet Inc. at 6.4%. Microsoft Corp., Amazon.com Inc., Broadcom Inc., Meta Platforms Inc., Tesla Inc. and Micron Technology Inc. round out the top 10, each with weightings between 1.6% and 4.2%.
The concentration risk extends well beyond equities. In the investment-grade bond market, AI infrastructure accounts for 49% of all net new issuance this year, while 87% of net new venture capital has flowed to AI companies, according to Apollo data. Even high-yield bonds show a 38% AI share of net new issuance, meaning investors who thought they held diversified portfolios may be far more exposed to a single theme than they realize.
AI's Grip on Capital Markets
The dominance of AI as an investment theme has made it nearly impossible for diversified investors to avoid. Foreign stock markets show similar patterns, with semiconductor names in Taiwan and South Korea dominating major emerging-market indexes, Slok said.
The data-center buildout alone is expected to drive about half of the 2% real GDP growth projected for the U.S. economy in 2026, according to Slok. That makes the AI theme not just a market story but a genuine macroeconomic risk with consequences for investors and consumers alike.
"The big risk is that the technology fails to deliver the hoped-for results," Slok said, pointing to the need for dramatic gains in worker productivity and corporate profit margins. So far, the only companies clearly profiting from AI are those making the semiconductors and equipment needed to power and operate data centers.
What's at Stake for the 493
The critical question, Slok said, is whether the productivity gains will spill over to the other 493 companies in the S&P 500 — those outside the trillion-dollar "Magnificent Seven" club. "There's no doubt that the Mag Seven have done well, but at the end of the day, is this going to spill over?" he said.
If the data-center buildout slows, the effect could ripple across the economy. A sharp decline in asset values could also hit consumer spending, as the wealth effect has played an increasingly important role in driving U.S. consumption since the pandemic.
For now, investor appetite for AI-related assets remains robust, according to Rob Haworth, a senior strategist at U.S. Bank Wealth Management. The S&P 500 finished about half a percentage point shy of its record close from early June, suggesting investors are rotating within equities rather than exiting the market entirely.
"Market sentiment is telling you the demand is there, credit spreads aren't widening out, so investors aren't being scared away by all of this debt issuance," Haworth said.
This article is for informational purposes only and does not constitute investment advice.
