Lisa Su just delivered the kind of quarter that makes Wall Street rewrite forecasts in red ink.
Q1 revenue: $10.3 billion, up 38% year over year. Adjusted EPS of $1.37, eight cents above the Street’s $1.29. And the headline number that mattered most: data center revenue surged 57% to $5.8 billion — beating consensus by $200 million and proving that AMD’s MI300 GPU and EPYC CPU franchises have become a serious counter-punch to Nvidia.
The Tape Got the Message Fast
AMD ripped 16% higher Wednesday, then tacked on another 5% Thursday after Morgan Stanley jacked its price target to $360, citing “strong data center momentum and AI chip demand.”
Q2 guidance came in at $11.2 billion — also above consensus. Server CPU market growth is now expected to exceed 35%, and AMD is closing in on a 1:1 CPU share ratio against Intel for the first time in the company’s history.
For perspective on the run: AMD is up 66% year-to-date and has more than tripled in the past 12 months. But it still trades at a meaningful discount to Nvidia on forward earnings — and that’s the setup that matters from here.
The Setup Into May 20
Nvidia reports Q1 fiscal 2027 earnings on May 20. The Street is pricing in $78 billion of revenue and 77% growth.
If NVDA delivers another monster, the entire AI semiconductor complex gets a second leg up — and AMD is the lower-multiple way to play it. If NVDA disappoints, AMD’s data center beat gives it a relative-strength buffer that pure-play Nvidia longs won’t have.
The bear case? Hyperscaler capex eventually rationalizes. Microsoft, Google, Amazon, and Meta are collectively expected to spend $530 billion on AI infrastructure in 2026. That number can’t grow forever — and when it stops, the entire AI semi trade gets a haircut at the same time.
The Play
AMD’s data center growth is now durable enough that pullbacks below the 50-day moving average are buyable into Nvidia earnings on May 20.
The asymmetric trade is the call spread, not the underlying. Implied volatility is elevated, but May 20 will move AMD by 10% in either direction. A May 22-expiration $230/$260 call spread costs less than a third of what owning the shares outright costs you in slippage if NVDA disappoints.
The structural trade: this is a multi-year AI capex super-cycle, and AMD has now proven it has a seat at that table. Drawdowns are buyable. Don’t chase strength.
Market Tea is investment research, not investment advice. Options trading involves substantial risk and is not appropriate for all investors. Always do your own due diligence.