Record highs and the great market debate

The S&P 500 has once again notched fresh record highs, surging more than 13 per cent year-to-date (YTD) and comfortably outpacing its historical average annual return. Driving the rally is a combination of robust earnings, apparently stabilising geopolitical friction, and the relentless build-out of artificial intelligence (AI) infrastructure.

This article was written on 11 August 2026.

Perhaps unusually, the new highs don’t appear to be accompanied by euphoria. Meanwhile, famed bears remain concerned. Market veterans like Michael Burry – who famously forecast the 2008 housing crash – have cautioned that equities may be forming a major top, and are at risk of a potential “1987-type fall.”

Every trade has a buyer and a seller, which of course confirms there are differing views. Either AI is giving birth to a once-in-a-generation productivity boom, or we’re witnessing an overleveraged bubble nearing its summit. Which is it? Here are the arguments..

The bull case

The optimistic thesis rests on measurable macro and microeconomic tailwinds. First, hopes of an agreement between Iran and Oman to reopen the Strait of Hormuz to commercial traffic have eased oil prices, tempering persistent inflation concerns. Second, the S&P 500’s Q2 earnings growth is expected to settle at 29 per cent year-over-year (YoY) more than double its 10-year average – even when excluding non-operating gains in SpaceX and Anthropic. Meanwhile, the cloud titans (think Microsoft Azure, Google Cloud, AWS) have reported acceleration in their collective revenue growth from 29 per cent last year to 54 per cent this year. Finally, as Ritholtz Wealth Management’s Josh Brown notes, a resilient market “takes out its own trash.” Recent liquidations by leveraged speculators – such as retail traders in South Korea and hedge funds on margin – are examples of the market shedding excessive froth while rewarding fundamental growth.

The bear case

The pessimistic view centres on a structural vulnerability in the artificial intelligence theme. While the bull case clearly dominates and is winning, the bear case rersts in the cracks.

Goldman Sachs, for example, estimates that AI investments account for nearly half of S&P 500 earnings growth, while Oxford Economics attributes a third of recent U.S. Gross Domestic Product (GDP) growth to AI. Capital expenditure on data centres, as a share of GDP, now exceeds the telecom build-out of the 2000s and the peak of the late 2000s housing boom. The counter to that of course is inflation over the last 26 years in the cost of building out any infrastructure.

While AI spending flows to semiconductors, memory makers, power infrastructure, and industrials, a disproportionate share of funding originates from a handful of private labs. Across Microsoft, Amazon, and Alphabet, aggregate analyst models indicate that over 70 per cent of hyperscaler AI sales do not yet come from broad enterprise adoption, but rather from compute spending and model resale concentrated in just two foundational startups: OpenAI and Anthropic.

If venture funding or revenue generation for these foundational labs falters, the entire capital expenditure (capex) waterfall risks an abrupt halt. As Apollo’s Torsten Slok points out, an expenditure cycle that adds 0.85 percentage points to GDP annually can unwind just as rapidly.

Another crack or fissure in the AI story relates to the pace of monetisation required for the hyperscalers to continue justifying their investment in the build-out of datacentres. If any disappointment in the rate of paid uptake by end consumers causes a hyperscaler to scale back their AI spend, some investors suggest the whole bubble could implode.

And then there’s Michael Burry. His thesis focuses on systematic leverage and 1987 analogies, extending beyond basic valuation concerns to focus on structural market mechanics:

“Remember, the market going up on falling volatility forces vol-targeting funds to leverage up, and brings leverage from other momentum strategies into play.” – Michael Burry, Cassandra Unchained

Burry is reportedly maintaining large short positions across semiconductor, hardware, and tech names – including Nvidia, Palantir, Micron, Tesla, Caterpillar, Applied Materials, and the iShares Semiconductor ETF (SOXX) – the latter recently collapsing as much as 29 per cent from its recent peak.

Burry’s primary concern is a self-reinforcing mechanical feedback loop: as stock prices rise and volatility subsides, systematic rules-based strategies (such as volatility-targeting and trend-following funds) automatically increase leverage. If market sentiment shifts unexpectedly, these algorithms can, and have, triggered forced selling.

It’s worth pointing out that while Burry reckons there are parallels to the Black Monday crash of 1987, that event prompted major stock exchange operators to introduce failsafes and guardrails that may prevent a repeat of the one-day falls almost 40 years ago.

Equity valuations and technology

A useful distinction when evaluating tech booms is separating stock market pricing from the ultimate economic utility of the underlying technology.

It’s one thing to point out an emerging technology is going to change the course of human history. There have been many such General Purpose Technologies (GPTs) in the past, from railroads to the car to television and the telephone, yet many investors lost their shirts by jumping in too late or paying too much. So, it’s quite another thing to assume all companies will win and investors in the theme will make off like bandits.

Technological transformations have historically involved over-investment, market consolidation, and business failures – much like the expansion of 19th-century railways or the early internet. A sharp market correction does not signify that AI has failed as a paradigm; it merely indicates that capital allocation moved faster than near-term monetisation.

Navigating the market horizon

The bulls and bears are presenting two contrasting realities: strong near-term earnings power supported by expanding profit margins, paired with historically high concentration and capital loop dependencies.

My view, which I have consistently presented, is that investors should diversify, taking some profits from the AI theme and investing those profits in uncorrelated asset classes and strategies.

Navigating this environment requires acknowledging volatility as a structural feature of innovation-led economies.

Whether the S&P 500 continues its upward trajectory or encounters a valuation reset, nobody knows. But by distinguishing between company fundamentals, systemic market mechanics, and long-term technological value, investors can rationally mitigate some of the risks.

For what it’s worth, I believe rebalancing is the key. Considering high-quality AA-rated private credit with no exposure to property developers, and zero-beta, pure alpha investment strategies such as arbitrage hedge funds makes sense, and may offer a destination for some of the profits taken from the record-breaking stock market.