The AI Bubble Question: S&P 500 Ascent Mirrors Pre-1987 Volatility
Table of Contents
A Familiar Climb
The trajectory of the S&P 500 over the last 18 months has begun to trigger alarms for market historians and quantitative analysts. Excluding the anomalous recovery periods following major recessions, the current pace of growth mimics a pattern not seen since the lead-up to the 1987 ‘Black Monday’ crash. The catalyst this time isn’t traditional industrial expansion, but a concentrated bet on Generative AI that has skewed the index’s performance to an unprecedented degree.
While the broader market appears healthy on the surface, a closer look reveals a precarious dependency on a handful of mega-cap entities. The “Magnificent Seven”—led by Nvidia and Microsoft—have acted as the primary engines of this ascent. When these entities are stripped away, the growth of the remaining 493 companies in the S&P 500 paints a much more stagnant picture, suggesting that the current rally is less of a general economic boom and more of a sector-specific frenzy.
The 1987 Parallel
To understand why the 1987 comparison is surfacing, one must look at the velocity of the rise rather than just the percentage. In the mid-80s, the market experienced a rapid, vertical ascent fueled by optimism about new financial instruments and corporate raiding. Similarly, the current market is riding the wave of Large Language Models (LLMs) and the infrastructure required to run them.
The danger in both eras is the decoupling of price from intrinsic value. In 1987, the market ignored warning signs of overheating until a systemic failure in portfolio insurance triggered a cascade of selling. Today, the risk lies in the “AI ROI gap.” While companies like Nvidia are reporting record-breaking revenues from H100 GPU sales, the companies buying those chips—the software startups and enterprise giants—have yet to demonstrate that AI is generating a proportional increase in their own bottom lines.
Concentration Risk and the ‘Halo Effect’
Industry analysts are increasingly concerned about the “Halo Effect,” where any company mentioning “AI” in an earnings call sees an immediate bump in stock price, regardless of the actual technical implementation. This has created a feedback loop: high stock prices allow these companies to hoard cash and acquire smaller AI startups, which in turn fuels the narrative of inevitable dominance, driving prices even higher.
This concentration is far more extreme now than it was in the late 80s. The S&P 500 is currently more top-heavy than it has been in decades. When a few stocks represent such a massive portion of the index, a single negative earnings report from a chipmaker or a regulatory shift in AI safety laws could trigger a broader market correction that drags down unrelated sectors.
The Fundamental Difference
There is, however, a critical difference between the current surge and the 1987 crash: the tangible nature of the hardware. In 1987, the growth was largely based on financial engineering. In 2024, the growth is backed by actual silicon. The demand for compute power is real, and the deployment of AI into coding, drug discovery, and logistics is providing measurable efficiency gains, even if those gains haven’t fully hit the quarterly balance sheets of the average S&P company.
The question remains whether the market has priced in a decade of perfection. If the transition from “AI hype” to “AI utility” stutters, the rapid climb of the S&P 500 may leave it vulnerable to a sharp, corrective descent, echoing the volatility of the late 20th century.
