I’m not chasing the Nvidia news. The reason is simple: what’s up today is sentiment, not compute capacity. A big bullish candle doesn’t mean a single additional GPU has actually been delivered, let alone that downstream demand has suddenly been revised upward. It’s mostly a recovery in risk appetite after the news came out. To me, headlines and market sentiment reflected in the chart are two separate things; you can’t conflate them.

Jensen Huang’s situation, frankly, has nothing to do with actual GPU production capacity. If you want to assess chip supply, look at TSMC’s production schedule and cloud providers’ capex—not at whose press release is better written. The monthly ramp-up in CoWoS advanced packaging capacity, the pace of HBM supply, and how TSMC allocates its N4/N5 capacity are what determine how many units Nvidia can deliver. Then it’s the capex guidance from Microsoft, Google, Amazon, and Meta that determines whether those cards end up in data centers or sitting in distribution channels. Over the past few quarters, whenever cloud providers have raised capex, orders across the supply chain have tightened. When capex has pulled back, even the most splashy product launch couldn’t support valuations. Press releases can be beautifully written, but production schedules don’t lie, and no one’s capex call is going to pick up the tab for them.

Today’s move in U.S. stocks looks more like the market treating “uncertainty resolved” as a bullish catalyst—a classic reversal trade, trying to pick a top against the news. The news itself may not be all that good, but as long as there’s no worse outcome, short-term money is willing to cover the shorts they had built up before. Then, as implied volatility falls on the options side and market-maker hedging flows give prices another push, the market can look stronger than the fundamentals warrant. I’ve seen this pattern plenty of times: stocks get hit before earnings, rally right after, then return to actual orders and guidance the next day. Chasing that first move means making money on sentiment—and losing it when sentiment flips.

Sentiment like this is most likely to produce a sharp wick when liquidity is thinnest, in the early hours of the morning. The order book is thin, spreads are wide, and stop-loss orders are clustered together, so one relatively large order can send prices into a long wick. Then the Asian session opens, liquidity returns, and prices pull back. I’m not taking that trade. It’s not that I don’t understand it; this kind of bet isn’t worth risking a position on. If I do get involved, I’ll wait for production schedules, capex, and order-book depth to point in the same direction instead of letting a wick at 3 a.m. pull me in.

I’ll save this call for now and come back tomorrow to check it. Let’s see whether my take that “sentiment is up, compute isn’t” lines up with the price action and supply-chain data.