Why do the biggest AI bets fall hardest when tech finally cracks?

When one group of assets races far ahead of the market, the size of that lead is itself a warning. Extreme outperformance tends to snap back, and the names carrying the heaviest bets fall hardest when the mood turns.

Tech just had its biggest 50-day run relative to the S&P 500 in recorded history. A six-sigma move, in plain terms: a stretch so large it should almost never happen if returns behaved normally. Then it cracked. On June 22, 2026, the heaviest spenders on AI led the way down, while a lighter name held up.

What a six-sigma move actually tells you

Sigma measures how far something strays from its average. One or two sigma is routine. Six sigma is the kind of reading that, on paper, should turn up about once in the lifetime of the universe.

So when tech's lead over the broad market hits that extreme, the lesson is not strength. A record lead measures how stretched expectations have become, not how durable they are. The rarer the run, the more it tends to revert. That reversion can mean tech falls, or it can mean the rest of the market catches up. Either way, the gap closes.

Not all tech is the same trade

Here is where it gets useful. When the selloff came, it did not hit evenly.

The names with the biggest AI capital bets fell hardest. Alphabet dropped about 6% on June 22, its worst day since May 2025, erasing more than $256 billion in value. Amazon fell nearly 5%. Microsoft and Meta each shed around 3%. Microsoft now sits roughly 30% below its all-time high.

Apple, the lighter spender on AI, held up far better. Its year-to-date decline is mild by comparison. When the AI story wobbles, the stocks pricing in the most AI upside have the most to give back. Heavy bets cut both ways.

The three times this pattern repeated

This is not new. Extreme leads have snapped back before, and the most-hyped names always took the worst of it.

• March 2000: tech towered over the broad market into the dot-com peak. The Nasdaq then fell roughly 78% over two and a half years, with the priciest names hit hardest.
• February 2018: an improbably calm, low-volatility market reversed in days. Short-volatility products were wiped out almost overnight.
• 2021 to 2022: high-multiple growth tech rolled over as rates rose. The longest-duration names fell 50 to 80%, while cash-generative mega-caps fell far less.

The honest other side

A rare run is not a timing signal. Mean reversion can take years, and it can simply mean the rest of the market broadens out rather than tech collapsing.

The strongest bull case: AI capex is building real future earnings. If the heavy spenders, Microsoft, Alphabet, Amazon and Meta, show that spending converting into faster revenue and fatter margins, the pullback was a healthy reset, not a top. Apple's own resilience has limits too. Its EU Siri regulatory issues mean the safe-haven read can break.

The evidence leans toward caution on the heaviest bets, not panic on the sector. Watch the upcoming mega-cap earnings and capex guidance. Watch whether the broad S&P 500 broadens out or whether tech leadership rolls over and drags the index with it. That split decides which story this was.