The divergence between mega-cap technology and semiconductor stocks suggests investors are becoming more selective about AI exposure. Strong revenue growth alone is insufficient when capital expenditure and valuations remain elevated.

My key indicators would be hyperscaler AI spending, semiconductor inventory levels, gross margins and the conversion of AI demand into sustainable profits.

I would also distinguish between semiconductor segments, as memory manufacturers face different market dynamics from chip designers and optical interconnect suppliers.

Ultimately, a correction becomes more concerning when earnings expectations deteriorate alongside share prices. Until then, I view the sell-off as a reassessment of risk rather than the end of the AI growth cycle.