Panmure Liberum strategist warns AI bubble could drag S&P 500 to 5,000

Joachim Klement thinks the AI party is running on borrowed time. The head of market strategy at Panmure Liberum is warning that the current AI investment boom looks like a bubble, and that it is likely to burst sometime between 2027 and 2028.

His target for the S&P 500 is 5,000 by the end of 2027. With the index recently trading around 7,723, that implies a drop of more than 35%.

The math behind the bearish call

Klement laid out the projection in his latest forecasts, published in October 2026. The pessimism is not limited to Wall Street: he expects European and UK equities to suffer as well.

His core argument is about cash. Free cash flow at the biggest hyperscalers is shrinking, according to Klement, while the debt costs tied to their capital spending keep climbing.

The spending in question is concentrated in a handful of US giants: Amazon, Microsoft, Alphabet, Meta, and Oracle. Panmure Liberum’s analysis from May 2026 put their combined 2026 capex at $658 billion.

To justify that outlay, Klement estimates those five firms would need to add $2-5 trillion in annual revenue.

Outside estimates point the same direction. Bloomberg Intelligence expects data-center spending to surpass $713 billion in 2026.

Bigger than the dotcom era

The October call builds on that May 2026 analysis, which argued the AI boom is already 60% larger than the late-1990s dotcom bubble. That study warned of severe downturns if tech spending retrenches.

AI-related market gains have reached $28.6 trillion since April 2025, a 31% increase in just six months.

That pace has only shown up three times before in modern market history: 1987, 1999, and 2009.

The Shiller P/E ratio sits near 40x, well above its historical average.

Klement has been consistent about where he stands. During an August 2026 appearance on CNBC, he described himself as “very bearish” on AI capex.

Scenarios: a pullback versus a full crash

A pullback of 4.5-6% in tech spending could send the S&P 500 down 20%, by his estimate. European and UK markets could fall more than 30% in the same scenario.

The more dramatic case is a full dotcom-style crash. In that scenario, Klement warns tech hardware stocks could plunge by more than 70%.

What this means for investors

The key variable to watch is the revenue gap. Klement’s thesis rests on capex continuing to outpace AI-generated income while borrowing costs rise. Quarterly earnings and capex guidance from Amazon, Microsoft, Alphabet, Meta, and Oracle become the scoreboard here.