Two weeks ago, Lisa Su walked onto AMD’s Q1 2026 earnings call and dropped numbers that should have triggered a national emergency on the “Nvidia has no competition” chat boards.
- Q1 revenue: $10.3 billion, up 38% YoY
- Data center revenue: $5.8 billion, up 57% YoY — now 59% of total revenue (was 38% in Q1 2024)
- AI accelerator (MI300 series): ~73% of data center revenue, or roughly $4.2 billion in a single quarter
- EPS: $1.37, up 43% YoY
- Q2 2026 guide: $11.2 billion (±$300M), up 46% YoY
- Stock reaction: +18% the day after the print
The stock has cooled since (the broader chip rotation hit on May 13), but the setup heading into NVDA’s Wednesday print is asymmetric. Here’s why.
The Meta deal already validated the second-source thesis.
Buried in AMD’s Q1 was a multi-year, 6-gigawatt AI infrastructure deal with Meta running on the upcoming MI400 series. That’s not a pilot. That’s the largest hyperscaler picking AMD as a second supplier for production-scale workloads. Microsoft, Oracle, and OpenAI are all in similar conversations.
Lisa Su’s pitch is simple: you can have two suppliers for your $200 billion data center build-out, or you can have one. The math on supply concentration only goes one way.
What Wednesday actually means for AMD.
Two scenarios — both work for AMD:
- NVDA beats but guides cautiously on supply. AMD rips. The MI355X is shipping now, the MI400 is on the road map for 2H 2026, and any whiff of Nvidia supply constraint sends every hyperscaler procurement team to AMD’s door.
- NVDA blows the doors off. AMD still rides the sector sentiment higher — chip rallies don’t leave the cheapest #2 behind.
Where it doesn’t work: NVDA prints in-line plus issues a soft guide. Then the entire chip complex sells off and AMD gets dragged down with it. That’s the binary risk — and based on Polymarket’s 95% beat odds, the market doesn’t think it’s likely.
The Play
The cleanest setup is to own AMD into Wednesday’s close, not after. Implied volatility is elevated across the chip complex but AMD’s options are cheaper than NVDA’s on a relative basis — the 30-delta call into Friday expiration is the simplest expression. If you’d rather avoid the binary, a stock position with a put hedge (the May 22 $145 put is the obvious one) keeps the upside open without the gap-down risk.
The structural read here is bigger than one earnings print: the AI capex cycle has moved into the “procurement teams want two suppliers” phase, and AMD is the only credible second supplier. That trade outlasts whatever happens Wednesday at 4:20 PM ET.
Stay caffeinated.