Can I make money from AI bots crawling my test site?

Probably not from the visits themselves. Crawlers aren't customers, and ad networks generally filter bot traffic rather than paying for it. They also won't buy products or generate affiliate sales just because they fetched a page.

The underlying data could be more valuable if you turn it into something people actually want to use. That's a different project, though, not a way to charge for the existing requests.

The infrastructure work sounds like the marketable part. A write-up about routing lots of subdomains through one controller, caching on a tiny server, and finding expensive queries could make a useful portfolio piece.

Include the limits and measurements you can actually verify. Request volume alone doesn't tell a reader how much work reached the VPS, especially when a cache sits in front of it.

I'd treat it as a useful but uncontrolled load test. Watch cache hit rates, origin requests, response times, database queries and resource usage during the traffic waves. That can show you where the bottlenecks are without pretending the crawler traffic is predictable or representative of real users.

First figure out what the traffic is costing you. Serve cached pages where possible and consider rate limits if the crawlers are crowding out useful requests. User agent strings alone aren't proof of who's making a request, so check how the claimed crawler can be verified before making special rules for it.

Robots rules can express your preference to cooperative crawlers, but they aren't a hard access control. If you need to protect the server, enforce limits at the edge or on the server too.