The problem with nginx commercial was it didn't work with the actual resources developers can bring to bare. If an nginx commercial license solves a problem, and the answer is "great, $5000 a year please" then the developer can't take that to their manager - the problem isn't critical yet, may not become critical, and they'll be asked to look for any other solution to work around it (I've been the guy doing this).
A per node, per core, per something license at a much lower rate would've saved them a lot of problems. Because $5 a month per node might get expensive quickly when scaling up, but thats a simple expense I can throw on the corporate credit card and then go "look, now the project is delivered".
A 1000x this. I see it with other software companies chasing enterprises (or at least VC backed companies) - Varnish Cache is a particularly egregious example.
In the end it hurts them because a developer or consultant may find a problem that is worth the license fee to solve - but with no hands on experience of the product due to the cost, it's hard to recommend it and take the reputational risk in case it doesn't live up to its promises.
Yes, they really miss the opportunity to get a foot into many smaller companies and scale up. For anything useful you need at least $10k/year (likely more), but what if you only need a tiny feature in addition to the open source software? Then this pricing feels just too much (although they of course have to financially support the open source software too) and it'll be very likely that you try to solve it with the open source nginx.
If they have modules - why not sell these modules and attach a "reasonable" price to it?
A per node, per core, per something license at a much lower rate would've saved them a lot of problems. Because $5 a month per node might get expensive quickly when scaling up, but thats a simple expense I can throw on the corporate credit card and then go "look, now the project is delivered".