The first alert came in on a Thursday morning, May 7, while I was walking the dog. NVIDIA calls, $195 strike, ten weeks out, about $72 million in premium at 4.3x volume over open interest - with the stock at $207.83. Deep in the money, long dated, enormous. Whoever bought those was not gambling on a headline; they were buying stock with a seatbelt.
I owned NVIDIA already, so there was nothing to do except not sell it. Which sounds like nothing, but it isn't: I had been nervous about my tech weight all spring, and that print was the reason I sat still instead of trimming the one name that was about to move.
Six sessions later
By May 14 the stock closed at $235.74. Thirteen percent in six sessions. The $195 calls had gone from big to enormous. My phone had buzzed a few more times on the way up - more calls, higher strikes, shorter dates, the tape chasing itself. That is usually the point where I start paying attention to the other side of the book.
The print at the top
On May 15, with NVDA sitting at $235.74, the alert that mattered arrived: $230 puts, June expiry, about $10.5 million in premium at 4.1x Vol/OI. Not panic-sized, but sober, dated, and struck just below the market. Someone was paying real money for downside protection at the exact level the stock had just reached.
I spent the evening reading. News first - nothing obvious. Then a Perplexity deep research on where NVIDIA's valuation sat versus its own history and versus the AI capex commentary that had been souring at the edges. The output didn't scream sell. It said: extended, crowded, and priced as if June couldn't disappoint.
So I did the boring thing. I sold the shares that the May rally had added above my target weight - the 13% the call buyer had handed me - and kept the core.
What June did
NVIDIA closed at $210.69 on June 18, the last session before those June puts settled. By June 25 it printed $192.53. From the May 15 top, that is an 18% slide. The put buyer was right, the late call chasers were wrong, and the difference between my May 7 self and my May 15 self was entirely made of two emails.
Same ticker, same month, opposite signals - and both were right in sequence. Flow isn't bullish or bearish. It's early. Your job is to notice which way it's early.
One tape walked me into the rally and the same tape told me when to take the gain off the table. If you hold crowded names, that second alert is the one you can't afford to miss. Put your tickers on watch - the tape will talk about them eventually.