Mike O'Rourke calls this a market flying on one AI engine, with NVIDIA defensible at 45x but Palantir dangerously stretched at 246x earnings. Target's comparable sales fell 3% and Ford absorbed $3 billion in headwinds, leaving both vulnerable to a double hit if AI-driven wealth evaporates. Lee Baker warns bond markets signal danger while equities ignore it, with the 10-year at 4.63% near a 12-month high and VIX near historic lows.
Don't wait: the analyst who called NVIDIA in 2010 just revealed his top 10 AI stocks. See the full list FREE now. Three market strategists appeared on CNBC on August 17, 2026, and one of them, Mike O'Rourke of JonesTrading, described the current tape in a way worth taking seriously.
He said the market feels like a plane flying on one engine, and that engine is AI. A very small group of companies tied to data center capital spending is carrying the index while the consumer economy is weakening. NVIDIA (NASDAQ:NVDA) trades at a $5.45 trillion market cap, and its Data Center segment brought in $75.246 billion in a single quarter.
Meanwhile, Target (NYSE:TGT) is guiding to approximately 2% net sales growth after a year in which comparable sales declined 2.5%. That divergence matters because if the engine holding the market up sputters, the ride down catches consumer names too, because household wealth and confidence are already stretched. O'Rourke said, "I think AI is a bubble.
That's my point of view, or at least not AI itself. It's the valuations of these stocks right now." NVIDIA at a P/E of 45x on trailing earnings that grew 85.2% year over year is defensible, and its $119 billion in total supply-related commitments suggests customers are still signing. What looks stretched is Palantir (NASDAQ:PLTR) at a P/E of 246x after 92.8% revenue growth, because that multiple prices an outcome no company has delivered at scale.
Meta Platforms (NASDAQ:META) trades at a P/E of 22x and has already fallen 10.49% year to date because its Q2 free cash flow collapsed to $784 million from $8.55 billion a year earlier, amid a $30.1 billion quarterly capex bill. A handful of extreme prices rest on the assumption that hyperscaler capex, running at what Jensen Huang called "the largest infrastructure expansion in human history", will continue to expand indefinitely. Any deceleration would compress those multiples fast. (See NVIDIA's Q1 FY2027 press release for the scale of commitments involved) O'Rourke's framing only requires the market's willingness to pay 246 times earnings for a piece of AI to fade, even if the underlying technology continues to succeed (riding a mania is fine as long as you plan the exit, and we walked through both halves in a free bubble survivor's handbook).
Extract — continue reading at the source.
