Ethereum + proof of stake + sharding. Right now every node on the Ethereum network validates every transaction using a similar proof of work system as Bitcoin. This creates a transaction bottleneck; I believe Bitcoin can do something like 7 TPS, Ethereum about 11, which is woefully inadequate for a global financial system (for comparison, I believe Visa does something like 20,000).
Proof-of-stake replaces this massive electricity waste with game theory. Every validator must put up a deposit (in ETH) to participate. If it cheats, it loses that deposit. The damage it can do to the network is limited to less than the deposit. Therefore, it's never profitable to cheat - and so they won't, or if they do, they will go bankrupt, remove themselves from the network, and self-limit the problem. And transaction validation can now be split among thousands of nodes (instead of having those thousands of nodes all validate the same PoW), which means the network can scale to tens or hundreds of thousands of transactions per second.
Bitcoin has a cultural bias to being conservative. Their selling point is that they were the first cryptocurrency, the Bitcoin network has never been hacked (even if individual exchanges or wallets have), and so they are a much more stable and dependable store of value. In the past they've been quite resistant to fundamental changes in how the blockchain works.
I'm pretty excited about Ethereum precisely because they're willing to experiment and rethink everything about their approach. I do suspect that soon after ETH2 is launched we will see a high-profile hack that steals everyone's ETH and requires a complete rollback of the network.
Proof-of-stake replaces this massive electricity waste with game theory. Every validator must put up a deposit (in ETH) to participate. If it cheats, it loses that deposit. The damage it can do to the network is limited to less than the deposit. Therefore, it's never profitable to cheat - and so they won't, or if they do, they will go bankrupt, remove themselves from the network, and self-limit the problem. And transaction validation can now be split among thousands of nodes (instead of having those thousands of nodes all validate the same PoW), which means the network can scale to tens or hundreds of thousands of transactions per second.