The chip trade is having one of those days where every bull’s favorite narrative and every bear’s favorite chart are both correct at the same time. On Tuesday’s CNBC Morning Call Sheet segment, Steve Grasso, Warren Pies, and Doug Boneparth all zeroed in on the same signal. Samsung posted blowout earnings yet its stock fell 7%, dragging the whole semiconductor sector lower.
Samsung is not US-listed and neither is SK Hynix, but the read-through hit everything from NVIDIA (NASDAQ:NVDA | NVDA Price Prediction) to Micron Technology (NASDAQ:MU). Micron fell 6.32% intraday. Advanced Micro Devices (NASDAQ:AMD) fell 6.12%. Even Qualcomm (NASDAQ:QCOM), which barely participates in the AI training cycle, slid 2.13%.
Why Great Earnings Still Sank the Stock
The panel’s argument is that the selloff was about margins and valuation exhaustion, not revenue weakness. Consider what “peak margins” actually looks like on the tape. Micron just reported GAAP gross margin of 84.6%, up from 37.7% the year prior, on revenue that grew 345.72% year over year. Management guided next quarter to $50 billion in revenue and roughly 86% gross margin. These are numbers you dream about at the top of a cycle, and that is exactly the problem. When a memory maker is earning 80-plus cents on every dollar of revenue, the market’s next question is always the same. Where does margin go from here? Not higher, usually. Grasso’s framing on the panel captures it. This is a commodity-style cycle, and you buy commodity semis at the trough or mid-cycle, not at 80% margins near the peak.
Valuation exhaustion is the technical name for that instinct. A stock can beat earnings, raise guidance, and still fall because the beat was already priced in. Reddit sentiment on Micron caught the tension in real time, with the dominant thread reading “I’m more confused by yesterday’s sell-off than the earnings.”
The Cycle Argument and the Summer Danger Zone
Warren Pies added the macro overlay. South Korean chip costs are up around 90 to 100% this year, mostly Samsung and SK Hynix, and trees don’t grow to the sky. Parabolic moves invite sharp pullbacks even when the earnings back them up. His seasonal tell is worth writing down. The momentum factor has been down each of the last five Julys, so July into August will likely be nasty as the market level-sets the big semi trade.
The prediction markets already agree. Polymarket puts the odds of NVIDIA closing above $200 on July 7 at just 2.2%, and the composite sentiment index on the stock has fallen 12.13 points over seven days. Micron is now trading well below its 50-day moving average of $852.09, with the stock down 20% over the past five days alone. The AI infrastructure buildout that Jensen Huang described in Nvidia’s Q1 filing as “the largest infrastructure expansion in human history” is still real. The long-term thesis staying intact and the near-term trade being dangerous are not contradictory positions. They coexist all the time.
How Not to Get Caught
Doug Boneparth’s advice was the practical one. Be very careful taking new positions at these prices; for those not yet in, dollar-cost average rather than chase, because FOMO bites hardest right before vicious short-term drawdowns. If you already own the AI winners, the question is whether you trim into strength. If you don’t own them, chasing a stock that has run 300.3% in the past year, as AMD has, means paying today’s price for tomorrow’s guidance. AMD’s trailing P/E sits at 207x, and its bearish Reddit signal peaked on July 1 with a post titled “+$75k on AMD puts.”
Averaging in across weeks or months protects against the exact scenario the panel described. You may miss the last leg of the melt-up, but you won’t buy the peak of a parabolic move on a Monday and wake up down 15% by Friday. Watch what NVIDIA does next. Data Center Networking grew 199% year over year last quarter. If July gets nasty and that number holds in August, the level-set becomes the buying opportunity.
Contact [email protected] for any questions or corrections.
Top Gaining Stocks
Top Losing Stocks