Ryan's correct, it is problematic but not impossible (if you temper your expectations). BBC R&D have been prototyping in this space for a while, it is feasible to have packet protection (i.e. encrypted payload) with a key shared by all - it is just incredibly weak security. Our multicast profile of HTTP over QUIC, based on the IETF version of the specification, adds stronger authenticity and integrity back at the application layer. We have an open source library,
nghq, and a toy sender and receiver that demonstrates some of that profile (just not the AEAD bit, which could be added with a bit more code).
To bring this back to Chromium, our earlier prototypes were based on libquic QUIC v33 and we were able to influence its behaviour from the outside. It's not easy but is feasible.