I2P claims to try to defend against large scale traffic analysis, but they are a underfunded project with few contributors. There was some mention of implementing cover traffic which would solve the issue (at the cost of massively increasing traffic), but I don't think that's happened yet.
I2P, being fully decentralised, is also very vulnerable to a sybil attack. Join thousands of nodes to the network, wait until you are strategically placed, then follow the traffic streams routed through your nodes.
Of course, sybil attacks are a concern in any open network. In theory the tor directory authorities are able to deny new nodes so they have some recourse, but in practice if you stagger your new nodes you can still infiltrate the network. :/
The fact is, anonymity systems are a hard and unsolved problem. That's not due to the source of the funding. We take what we get.
