Is the idea to push customers towards Ethernet services?
RevK put the reason on his blog.
They pay a ton of money to BT for backhaul, to support 330mbit services would require an extra 10s of thousands per month in expense to support the service, as essentially burst speed costs money (which many in the UK wrongfully thought burst speed is free based solely on costs of DSL ports in the exchange).
I think this was put on a ispreview news article as well about why a aaisp 330mbit trialist was not getting the full 330mbit speeds.
So e.g. if you paying BTw £50 per mbit of capacity, you need to add say 300mbit to support the service then thats 300 x 50 = 15000gbp per month to support the service and thats assuming on average two users wont try to burst to full speed at the same time, as aaisp have a policy of no visible contention.
I think with the BT pricing, to get their lowest price you have to pre commit, so an isp pays for the bandwidth "just in case" it gets utilised. So AAISP would be assuming it may get utilised even if all their 330mbit customers are grannies just checking emails.
I am speculating aaisp dont anticipate high sales, so e.g. if they predicted 5 customers on g.fast their revenue would be 5 x the consumer cost of the product, and their costs would be 5 x the port cost to openreach plus the 15k or so to BTw.