Further proof it is not me but Google Fiber infrastructure: I changed my MAC address to a new one and got a fresh prefix delegation: 2604:5500:706b:3e00::/56 via DHCPv6-PD. This prefix is ALSO not routed to me. I also received the address 2604:5500:706b:0:f09b:7d59:fba6:205 via SLAAC. This address works and *is* therefore routable, so SLAAC addresses from the headend are working, meaning it is only DHCPv6-PD prefixes that are not. I have a dedicated server here in Seattle at Vultr. It has an IPv6 address of 2001:19f0:8001:704:5400:5ff:fe8c:fac5/64, which I am testing IPv6 connectivity by sending traffic to and from it to my router at home. Again, traffic TO and FROM the SLAAC address is working. For example, here is an mtr from the vultr machine: vultr:~# mtr -z -n --report-wide -c 10 2604:5500:706b:0:f09b:7d59:fba6:205 Start: 2025-08-13T07:02:07+0000 HOST: vultr Loss% Snt Last Avg Best Wrst StDev 1. AS??? 2001:19f0:ffff::1 0.0% 10 0.2 0.2 0.2 0.4 0.1 2. AS??? 2001:19f0:fc01:a::6464:6401 0.0% 10 1.0 0.7 0.5 1.0 0.2 3. AS??? 2001:19f0:fc00::a42:1d9 0.0% 10 3.8 2.7 1.1 9.9 3.0 4. AS??? ??? 100.0 10 0.0 0.0 0.0 0.0 0.0 5. AS??? 2001:19f0:fc00::a42:1f5 0.0% 10 1.0 4.8 1.0 20.0 6.1 6. AS??? 2001:19f0:fc00::a42:1e9 0.0% 10 1.9 8.0 1.4 38.5 13.4 7. AS??? 2001:504:16::4add 0.0% 10 1.1 1.1 1.1 1.2 0.0 8. AS19165 2604:5500:7003:2222::2 0.0% 10 1.4 3.0 1.2 12.3 3.5 9. AS19165 2604:5500:7008:10::1 0.0% 10 2.5 11.6 2.1 89.1 27.3 10. AS19165 2604:5500:70b9::1 0.0% 10 2.4 14.8 2.2 125.7 39.0 11. AS19165 2604:5500:706c::1 0.0% 10 24.1 8.8 1.7 48.0 15.4 12. AS19165 2604:5500:706b:0:f09b:7d59:fba6:205 0.0% 10 1.4 1.5 1.4 2.2 0.2 Your headend is hop 11, 2604:5500:706c::1. The return path back to my machine at vultr is below. Your headend in this case is hop 1, 2604:5500:706b::1. Start: 2025-08-13T00:10:44-0700 HOST: thistle Loss% Snt Last Avg Best Wrst StDev 1. AS19165 2604:5500:706b::1 0.0% 10 0.8 1.0 0.7 2.3 0.5 2. AS19165 2604:5500:70b9::1 0.0% 10 68.8 8.6 1.2 68.8 21.2 3. AS19165 2604:5500:7008:10::1 0.0% 10 1.4 5.3 1.2 27.7 8.2 4. AS19165 2604:5500:7003:2222::2 0.0% 10 0.9 2.1 0.9 8.9 2.5 5. AS19165 2604:5500:7003:2222::1 0.0% 10 0.6 0.7 0.6 0.7 0.0 6. AS??? 2001:504:16::4ff9 0.0% 10 1.4 2.2 0.7 10.6 3.0 7. AS??? ??? 100.0 10 0.0 0.0 0.0 0.0 0.0 8. AS??? 2001:19f0:fc00::a42:296 0.0% 10 11.9 5.6 1.8 17.8 6.1 9. AS??? 2001:19f0:fc00::a42:1e6 0.0% 10 1.1 1.1 1.0 1.2 0.0 10. AS20473 2001:19f0:8000:5e::33 0.0% 10 1.0 1.1 1.0 1.1 0.0 11. AS20473 2001:19f0:8001:704:5400:5ff:fe8c:fac5 0.0% 10 1.1 1.2 1.1 1.2 0.0 The other address configured on my router at home is 2604:5500:706b:3e00::1/64, which is inside the delegated 2604:5500:706b:3e00::/56 subnet. Traffic to and from this IP does NOT work, the headend does not have any route to 2604:5500:706b:3e00::/56 and considers the prefix to be a bogon. This is proven by trying to send traffic to and from the 2604:5500:706b:3e00::1 address mentioned above. From vultr we get: vultr:~# mtr -z -n --report-wide -c 10 2604:5500:706b:3e00::1 Start: 2025-08-13T07:16:48+0000 HOST: vultr Loss% Snt Last Avg Best Wrst StDev 1. AS??? 2001:19f0:ffff::1 0.0% 10 0.3 0.2 0.2 0.4 0.1 2. AS??? 2001:19f0:fc01:a::6464:6401 0.0% 10 0.6 0.5 0.4 0.6 0.0 3. AS??? 2001:19f0:fc00::a42:1dd 0.0% 10 3.2 2.1 1.1 5.0 1.4 4. AS??? ??? 100.0 10 0.0 0.0 0.0 0.0 0.0 5. AS??? 2001:19f0:fc00::a42:1f5 0.0% 10 1.1 2.9 1.0 15.4 4.4 6. AS??? 2001:19f0:fc00::a42:1e9 0.0% 10 6.4 6.4 1.1 12.5 3.7 7. AS??? 2001:504:16::4add 0.0% 10 1.1 1.1 1.0 1.2 0.1 8. AS19165 2604:5500:7003:2222::2 0.0% 10 1.4 1.4 1.3 2.0 0.2 9. AS19165 2604:5500:7008:10::1 0.0% 10 138.7 31.6 2.4 143.4 57.8 10. AS19165 2604:5500:70b9::1 0.0% 10 3.1 5.4 2.2 24.8 7.1 11. AS19165 2604:5500:706c::1 0.0% 10 1.7 2.4 1.7 8.0 2.0 12. AS19165 2604:5500:706b:0:a62b:8cff:fe8b:120b 0.0% 10 1.6 1.6 1.5 1.8 0.1 13. AS??? ??? 100.0 10 0.0 0.0 0.0 0.0 0.0 Hop 12 is a different customer: 2604:5500:706b:0:a62b:8cff:fe8b:120b, as this is an SLAAC address we can work backwards and determine their MAC address, which is a4:2b:8c:8b:12:0b. Again: my router has 18:e8:29:bc:45:2e as its MAC address and is using 2604:5500:706b:0:f09b:7d59:fba6:205/64 as its IPv6 endpoint address. The router at hop 12 belongs to a DIFFERENT customer. Your DHCPv6-PD server is handing out duplicate prefixes. This prefix is in use by another customer (and this can be seen by the fact that the customer's router appears in the traceroute), which explains why it isn't being routed to me (the rejection of packets along the return path is the right thing to do here, it means you guys do at least implement BCP38 correctly) but your DHCPv6-PD server assigned it to me. I am frustrated, because I have spent this entire day being told it absolutely couldn't possibly be Google Fiber, but here I am with a prefix handed to me that is already assigned to another customer, who is *actively using it*. Please advise on when this issue will be fixed. If I do not get an appropriate response, I will be forced to blog about this issue, as I believe this to be a legitimate security concern at this point.