AMD’s Q3 earnings should have sent the stock higher in any normal environment: revenue of $13.38 billion, up 45% year over year and a record; data center revenue of $8.55 billion, up 97% year over year and nearly doubling; Q4 guidance of $17 billion, far above analysts’ $16.3 billion expectation; and Lisa Su even said on the earnings call that server revenue would double year over year in 2027. Shares briefly rose more than 7% after the report.

Then Thursday came, and AMD fell 17%.

This disconnect between “earnings beat expectations” and “shares plunged” is precisely the best entry point for understanding the nature of this sell-off. The market isn’t pricing AMD’s Q3 results; it’s repricing the “sustainability of AI capital spending.” Doubling data center revenue was an unambiguously positive development before the midterm elections; after the election, with Democrats signaling tighter regulation, data center approvals could slow, turning the same doubling figure from “proof of growth” into a “signal of a cyclical peak.” Combined with mechanical selling from hedge fund deleveraging, AMD, as a high-beta proxy for AI hardware, was hit much harder than the broader market.

CLSA maintained its “Outperform” rating and $710 price target after the earnings report, citing a “strong product cycle ahead.” I agree with half of that argument: the product cycle is real, but it will only translate into revenue if data centers are built on schedule. At a time when regulatory visibility is so low, AMD’s earnings visibility is actually more dependent on policy variables than Nvidia’s. I’ll keep AMD on my watchlist and wait for the narrative around data center approvals to stabilize before deciding whether the “earnings beat” is a starting point or just a final flicker. #Gate股票畅聊季