Cloudflare beat and jumped 15%. Peloton beat and fell 13%. Both reported the same night.

Four companies beat this week. Two went up and two went down, and I don't think the beats explain any of it.

Peloton reported Wednesday night. Revenue $608M against about $598M expected, EPS 13 cents which is exactly what the street had. FY26 was the first annual profit in the company's history. The stock fell about 13% on Thursday. Cloudflare reported the same night. Revenue $696.1M, up 36%, against roughly $664M. Adjusted EPS 29 cents against 27. The stock jumped about 15%. Same setup, opposite reaction. The difference is in what each of them said about next year.

Cloudflare raised the full year on both lines. Revenue guidance went to $2.86-2.87B when the street was at $2.81B, and EPS to $1.25-1.26 against $1.21. Q3 guided to $736-737M, up 31%.

Peloton guided FY27 to $2.3-2.4B. The street wanted $2.42B. That is a revenue decline of about 4% from a company that just printed the first profitable year it has ever had.

So one of them beat and said next year is better than you thought, and the other beat and said next year is smaller. The quarter that already happened was the same kind of quarter in both.

The two big prints this week are the same story with an extra step in it.

AMD beat revenue, beat EPS, and guided Q3 to about $13B when the street was at $12.5B. It closed down 7% on Wednesday. The line that got read was in the guide rather than the print: gross margin guided flat at about 56%, into a quarter they are guiding up 13% sequentially. Jean Hu said on the call that the data center AI mix runs slightly below the corporate average, so the part growing fastest is also the part holding the margin flat.

SpaceX beat on revenue, $7.8B and up 92%, with all three segments ahead of estimates. It closed at $108.27 on Wednesday, down 13.6%. Capex for the quarter was $18.37B, and $15.8B of that went to AI infrastructure, up from $7.7B in Q1. Connectivity was the only segment that made an operating profit. The AI segment did $2.56B of revenue and lost $1.26B on it. I should be fair to SpaceX and say the drop is not purely the capex. There was a lockup expiry the next day, so some of that move is just supply.

What I take from the week is that a beat is a fact about a quarter that already closed, while the price is a bet on the next several. When those two disagree you get Peloton. When they agree you get Cloudflare. AMD and SpaceX are the third case, where the forward revenue number is fine and what it costs to get there is the thing that moved the stock.

Question for anyone who watches a lot of these. Do you have a rule for telling a real raised guide from management pulling demand forward to defend a multiple? I have never found a clean one, and Cloudflare growing 36% with the stock at a record is exactly the setup where I would want one.

(Numbers from each company's Q2 release and the Aug 5-6 coverage. Information, not advice, so tell me where I have this wrong.)