Mist wolf, I have had that problem and fixed it with a heavier buffer. The bolt would attempt to extract the round but since pressure was to high in the chamber it would not fully extract and was still there on bolt closing. Also replacing the spring with a stronger one made it worse. It may be over gassed as well. In either case slowing the bolt will help. The best way is a heavier buffer. If the spring is too strong that will also cause short stroking and the bolt to close too fast. That has been my experience. If you have a better idea let's hear it and why.
If the carrier is moving so fast, the empty hasn't enough time to make it out of the ejection port, you'll know it. Long before the carrier reaches those kinds of velocities, the carrier will be hitting the hammer back so hard, a sharp, stinging slap will be felt in the trigger bow. I can tell you from experience, it will be a very painful stinging slap.
When I built a shorty AR with a 10.5 inch barrel, a carbine weight buffer was used. Because the gas port was .083 inches, it was over-gassed and carrier speeds were too fast- fast enough that I felt a painful slap through the trigger bow. I did not experience any failures to extract or eject but it did send the brass into the next area code.
The carrier only has to travel about halfway back until the mouth of an empty will clear the leading edge of the ejection port. That gives the empty plenty of time to eject. Faster carrier speeds do not cause ejection problems, it only uncovers existing or impending problems. One thing that must be kept in mind when troubleshooting- Failures don't always just suddenly start. Springs don't always just lose tension all at once. Extractors wear over time. Rifles don't usually work fine until they suddenly start malfunctioning.
An AR will start to get an occasional malfunction, but the malfunction gets dismissed as a fluke. In reality, it's an indication that something has reached the end of it's useful life and is on its way to failure.
Extractors & ejectors and their spring don't wear away all at once. They may wear to the the point that they cause malfunctions when carrier speeds are on the high side, but work fine when carrier speeds are slowed down. Adding a heavier buffer or lighter spring may drop speeds enough to allow a worn extractor or ejector to work reliably again, but the wear is there and will continue and more malfunctions will occur.
It's human nature to think that because we change something (buffer weight, action spring rate) that makes a problem go away (empty cases again are completely ejected), the problem was caused by something else (carrier speed was too fast to allow empty case to clear ejection port) when what it really does is mask the true problem (weak ejection and/or extraction)