This is only for B2C though.
It's all cost for me, none for them, just to make them behave decently in the first place.
The competitors could not respond with counter comparisons due to the first company's license, and so instead they added their own prohibitions to stop the first company.
Some of them just adopted the same blanket prohibition, but some were more subtle, going for reciprocity rather than prohibition--you could not publish benchmarks comparing their product to others unless you could grant or obtain for them permission to also publish benchmarks doing the same comparison. Some also required that you publish complete details of your testing so that it could be reproduced.
I don't know if anyone still uses the reciprocity approach. I'd expect that there would be plenty of loopholes that someone (Oracle?) could use to get around it that any big player using it would have given up and went for a straight prohibition.
Actually there is a good story behind it. See: "Larry Ellison allegedly tried to have a professor fired for benchmarking Oracle" [0] and the related HN discussion. [1]
[0] https://danluu.com/anon-benchmark/ [1] https://news.ycombinator.com/item?id=15886333
"e.Benchmark Testing. You may not disclose the results of any benchmark test of either the Server Software or Client Software to any third party without Microsoft’s prior written approval."
Hasty addition: I doubt random benchmarks would be threatened with legal action -- but the key is you are technically breaking your license agreement by posting a benchmark. I'm sure the likelihood of whatever DB company pursuing action depends on both the company, who you are, and in what avenue you posted your benchmark.
If you're well known, have a large audience, and publish a bad benchmark of Oracle DB, I would be very surprised if Oracle didn't contact you directly and demand some action.