Racing on smooth roads isn't an activity which is particularly gentle on bike frames. Crashes are just a statistical fact of life, and people really go at it on races - sprints, climbing out of the saddle, it's hard on the bike. Last proper road race I've been on, one of the people guarding the roads messed up and started signalling we were to turn right when we were supposed to go straight, with the bunch going at around 45 km/hr. So people in the front start braking and turning, wheels touch and next moment I see out of the corner of my eye a beautiful red CF Willier just soaring through the sky, doing a summersault - I think everyone in that pile-up finished the race on their bikes, though. Last year I did that same race, I was at the front of a group of dropouts from the first group and a car veered into the course, I braked and tried to veer off sideways but couldn't avoid crashing into it. Got up after a minute, gave the bike a look-over, found that it was okay and everything spun properly and didn't make sounds, hopped back on and went riding (then had to stop a kilometer away and lie down a bit due to knee pain before finally continuing). CF is tough - not indestructible and not forever, no material is, but it's tough. Steel has an outsized reputation amongst retrogrouches.

Here's an old, old article where frames were fatigue tested. While the forces are higher than normal for a rider climbing out of the saddle (cca 750N for a relatively light rider putting out a mean torque of cca 85Nm), they are significantly below maximum achievable on a bike (for instance, if you do accelerations into a sprint from a standing as part of your training, forces over 2.5 times the cyclist body mass were measured - about 1900N).
12 High-End Frames in the EFBe Fatigue Test (sheldonbrown.com)