The Highest-Probability Way to Trade NVDA Earnings Tonight?

NVDA reports after the bell tonight, and this is the single most important event of the week for the entire technology complex, for AI capex sentiment, and potentially for sector rotation across the market! If NVDA re-prices the AI narrative, QQQ moves, growth vs value rotation shifts, even energy and cyclicals feel it through flows.

Now here’s what’s interesting. Consensus is extreme: 65-66B revenue, +66% YoY. EPS up 70%, Data Center nearly the whole engine. The base case is already beat and strong guide. You'd expect options to price this like a bomb, but they aren't!

The at-the-money straddle implies roughly a 5-6% move. Over the last 12 quarters, the average implied move was closer to 7.5%. By NVDA's own standards, this event is being priced smaller than usual. That's the first non-obvious signal.

The second one is even more important. Historically, NVDA's implied earnings move trades at about 1.5x the tech sector (XLK). This quarter, that ratio is closer to 0.9x. Read that again: the market is pricing NVDA as less idiosyncratic than the sector, at a moment when AI capex concentration arguably makes it more idiosyncratic than ever.

Yes, front-week IV is high (72% vs 55% baseline). Yes, there will likely be IV crush. But the lazy trade "short the rich IV" assumes the event premium itself is bloated. This time, the event premium is compressed relative to history and relative to tech. That changes the game, so we're not putting on calendar spreads today.

The edge, in my view, sits in:

NVDA vs sector variance
Defined-risk or asymmetric volatility harvesting

The market is not overpricing fear, it's compressing NVDA's uniqueness into sector volatility, and if that assumption breaks tonight, the move won't care about your straddle math.

Because the event move is priced relatively small (5-6%) and the front-week premium is not unusually fat versus history, forcing a weekly Jade Lizard would mean selling compressed event variance with thin margin for error. So instead of playing the binary print, we step out to April (51 DTE) and build a safer Earnings Jade Lizard.

My structure (April 17 expiration, 51 DTE): Sell 170 Put @ 3.95, Sell 195 Call @ 15.05, Buy 200 Call @ 12.55. Net credit: $655 per 1-lot, Max profit: $655, Probability of Profit: 83% (!), Theta: +$9.5/day.

Why this works in this specific NVDA setup:

You're not relying on IV crush alone. We're selling premium in a tenor where IV is elevated relative to realized, but not purely event-inflated. That gives us better structural edge than selling the weekly hump.
No upside risk. The call spread width (5 points) is fully covered by total credit collected. If NVDA rips on earnings, you don't care.
Downside positioned outside implied band. The 170 short put sits well beyond the implied move zone.
Positive convexity to volatility normalization. Post-earnings, the front hump collapses, and term structure typically flattens. April volatility will likely drift lower from 49-50% toward mid-40s as event uncertainty clears. That benefits the Jade Lizard without needing a directional win.

Given the relative compression of NVDA event variance versus sector, I prefer collecting structured April premium over gambling on weekly IV crush. We're trading volatility geometry.

Next step: management plan if earnings break the implied range.

0 comments