IPV6 can have speeds up to 10 gigabytes per second and have a trillion times more internet addresses for new types of applications.
Clarification (H/T to commenters) –
The next-generation IPV6 network is faster because of the underlying physical infrastructure, wires/fibers and hardware. The speed is not related to IPV4 vs IPV6. Even faster networking equipment is commercially available today and will work with IPV4 just fine.
There are 96 more bits of address space in IPV6. That’s 64 billion trillion times the address space.
There is an online site for performing IPV6 vs IPV4 speed tests. They seem to be indicating that IPV6 is 70% of the speed of IPV4 for the downstream part but can be 6 times faster in places like South Korea.
There is nothing inherently faster about IPv6 either. In fact, it is slightly slower due to the increased header overhead. The processing overhead of sending 1500 byte packets over a 10gbps link can be significant, especially when more of those 1500 bytes are used in the headers and less in payload. Jumbo frames should be supported reducing computational and space overhead. 10gbps can also wrap a TCP sequence number (32 bits = 4GB) in 3.2 seconds, potentially leading to ambiguous out-of-order retransmissions. Many TCP implementations will have to support PAWS to operate at 10 gbps. Another issue is that very few buses operate at 10 gbps, such as SATA III at 4.2 gbps.