A Shifting AI Landscape
Nvidia continues to occupy a central role in the artificial intelligence buildout, but the dynamics around the stock are quietly shifting. Over the past month, shares climbed roughly 10%, a respectable gain that nonetheless trailed several of its peers. AMD has surged about 60% over the same window, while Intel has effectively doubled. That relative underperformance is not a sign of weakness in Nvidia's fundamentals so much as a reflection of where the market believes the next phase of AI growth will unfold.
The first chapters of the AI story were almost entirely about training—building the massive compute infrastructure required to teach large models. Nvidia's GPUs were the indispensable engine of that effort. We are now entering what looks like the middle innings of the cycle, where demand is migrating toward inference, the deployment phase where trained models are put to work serving real users at scale. Inference workloads can be handled efficiently by a broader range of silicon, including CPUs and diversified chips. That is precisely why competitors with strength outside of pure GPU dominance have been catching investor attention.
Competition Intensifies, but Nvidia Is Not Standing Still
The competitive picture has grown more complex. Hyperscalers are increasingly designing their own chips—Google's TPU is the most prominent example—optimized specifically for inference workloads in their own data centers. This vertical integration represents a structural challenge that Wall Street is clearly pricing in.
Yet Nvidia is hardly resting. The company continues to expand beyond chips themselves and into the broader ecosystem required for next-generation AI infrastructure. A newly announced partnership with Corning aims to scale fiber and optical connectivity, a critical and often overlooked piece of modern AI data center design. The deal sent Corning sharply higher and underscored that Nvidia intends to shape every layer of the AI stack, not just the accelerator card.
There is also a roadmap argument to be made. Nvidia's GPUs still drive roughly 90% of the most popular AI workloads. Historically, when AMD has begun to close the gap, Nvidia has answered with a generational leap that re-establishes its lead. The upcoming Vera Rubin chip sits on the horizon as the next such inflection point. Whether it once again pulls the company decisively ahead is uncertain, but the pattern is suggestive.
The Numbers Behind the Narrative
The fundamentals remain extraordinary. Wall Street is looking for roughly $86.6 billion in second-quarter revenue, representing about 85% year-over-year growth. Earnings growth in the 60% to 70% range is expected to accompany that top-line expansion. By any traditional measure these are video-game-level numbers, and the valuation, set against this growth, still looks reasonable rather than stretched.
Technically, the stock rallied to around 216 before pulling back to 198, where it bounced cleanly off its 21-day exponential moving average. If momentum continues to build, a renewed challenge of the prior highs is plausible. With earnings approaching, however, the case for a definitively bearish posture is thin—the bull-versus-bear debate on this name currently feels closer to bull-versus-raging-bull.
A Broken Wing Butterfly for a Measured Bullish View
For traders who want exposure through earnings without taking on full directional risk, a broken wing butterfly in the May 22nd expiration captures the event itself. The structure involves buying the 205 strike, selling two of the 215 strikes, and buying one 220 strike. That creates a $10 wide spread on the lower side and a $5 wide spread on the upper side—an asymmetric butterfly that skews the risk/reward profile in the trader's favor.
The ideal outcome is for the stock to drift up to the 215 short strike. After earnings, implied volatility collapses, and because the short strikes sit closer to the money, their volatility crush is sharper than at the 205 or 220 wings. That dynamic is what powers the trade. The strategy assumes a roughly 4% to 5% move higher from current levels rather than a dramatic surge.
The risk is well known to anyone who has traded butterflies: getting the direction right but seeing the move blow past the short strike, watching profits evaporate as the stock keeps running. AMD recently moved twice its expected move on earnings, a useful reminder of how violently these names can travel. The asymmetric construction—long $10, short $5—mitigates this somewhat by keeping the extended profit curve more forgiving on the upside than a symmetric butterfly would. Total risk remains capped at the debit paid.
For those expecting a larger move, the structure is easily adjusted. Strikes can be pushed out to a 210/220/225 configuration, for example, allowing more room for upside while maintaining the same architectural logic.
A Call Diagonal for Flexibility and Time
A more aggressive alternative is a call diagonal that extends beyond the earnings week itself. The setup involves buying the May 29th 205 call (after earnings) and selling the current week's 210 call, with two days remaining until that short expires. The net debit comes in around $8.75 for what is structurally a $5 wide diagonal—more than the width of the strikes—because the longer-dated implied volatility is elevated and the trader is paying for time.
This trade requires active management. Until Friday, there is genuine risk of loss if the stock screams through the short strike before the near-term option expires. After that, the position can be rolled: extending duration, moving the short strike further out, and progressively chipping away at the net debit. By the time the earnings expiration is in play, the credits collected from rolling can substantially reduce or even neutralize the original cost.
The real value of the diagonal is flexibility. If Nvidia begins to grind higher, the 210 short call can be rolled up to 215, then 220, ideally for credits as long as the move is not too sharp. Each successful roll embeds a vertical spread inside the position while reducing cost basis. As the earnings event approaches, the trader retains the option to maintain a bullish stance or unwind. With enough rolls, what began as a pricey diagonal can transform into a substantially lower-cost play on the event itself.
Conclusion
Nvidia sits at an interesting inflection. It remains the dominant force in AI compute, with eye-watering revenue and earnings growth and a roadmap aimed at preserving its technical lead. At the same time, the AI cycle is broadening from training to inference, hyperscaler-designed silicon is creeping into share, and competition is real and accelerating. For options traders, that combination of strong fundamentals, building volatility, and credible competitive pressure creates rich terrain. A broken wing butterfly offers a defined-risk way to lean bullish into earnings while harvesting volatility crush, while a call diagonal trades a higher upfront cost for the kind of flexibility that lets a trader adapt as the stock and the broader AI narrative continue to evolve.