shallow knowledge on my side, but it seems that the AT proto requires some fancy relay node that, as of now, are only hosted by big corps, so no one has a a fully independent bsky instance
This seems to be left out of the conversation when comparing Bluesky with the wider fediverse. The AT protocol relies on a bespoke setup and is complicated to implement. ActivityPub is comparatively simple, and can be implemented on pretty every thing with just a few endpoints.
There is not a bespoke setup that you need to implement atproto. In fact, there are already a variety of applications making use of it (some to a higher degree than others). There are community implementations of app views, relays, PLC directories, and PDSes already in the wild, and - although I admittedly have a biased ear on the conversation - developers tend to appreciate the _lack_ of complication when implementing things.
you can run a resequencing relay good enough to feed an atproto appview pretty cheaply - it just needs to subscribe to a few thousand websocket endpoints to get a live tail of the whole network (incoming traffic is well under 50Mbps at peak in my experience)
running an archival mirroring relay is storage-intensive (on the order of tens of TB iirc?) but only serves as an optimization (you can backfill full atproto repositories straight from the relay instead of needing to reach out to the relevant data server)