If The Market Doesn't Broaden, The AI Trade Could Crush It

Adam Taggart | Thoughtful Money® Jul 17, 2026

Adam Taggart's takeaways

Top takeaway He frames the episode around one risk: Wall Street is pricing the AI complex for better than perfection, so anything that forces analysts to write down forward earnings estimates would hit far beyond technology stocks.
He is not calling an imminent AI collapse; his concern is that leadership has become too narrow, and the next sustainable move higher needs something more than just chips.
He points to early but weak signs of broadening — regional and large banks breaking out, biotech and healthcare breaking out, and some strength in homebuilders — as the test the market has to pass.
He reads the index itself as stalled and undecided as of the 2026-07-17 recording: the S&P 500 sitting around 7,500–7,600 for a few months in a large diamond/triangle formation, roughly 60 points below an all-time high, looking like it wants to break higher.
He flags bubble signs elsewhere that are pressing up against the ceiling without lifting the index — SK Hynix showing signs of a bubble over the past week, memory and disk makers up sharply, and SpaceX falling back toward its IPO range after running above 200.

Top takeaway He is introduced as a co-guest from New Harbor interviewed on whether the market's dependence on the AI trade is now the dominant risk, though the transcript attributes no individual view to him.

Top takeaway His top point is that if the AI trade fell apart, the whole market would come down with it, because the market is too narrow to absorb an AI-led correction.
He argues the concentration is extreme: the AI sector is almost half the market cap of the S&P 500 and almost three quarters of the Nasdaq 100, so the financial markets are all in on this one trade.
He links the AI complex to the wider economy, arguing that much of US GDP growth comes from this capex spending, so a material correction in the sector would carry a big negative impact beyond the AI stocks themselves.
He does not expect the theme to end soon — the party could go on for another three-plus years — but he says earnings expectations and valuations are off the map and that AI needs to hold its ground and then broaden into other sectors for the market to keep going.
He wants that broadening to happen relatively soon or the market will show signs of exhaustion, and he judges the current early signals — banks, biotech and healthcare, and some homebuilders — as still weak and incomplete.