If we would like to pick nits, we could say that the Visa transactions are not comparable to the Solana transactions because Solana transactions let users run complex programs that Visa does not, or that Solana transactions are not comparable to Visa transactions because nobody accepts SPL USDC and everyone pays out the nose to accept Visa, or that 2700 is a lot less than 6500, or that the 6500 number is misleading because the load should have peaks, and do we really know how much peak load Solana can handle, or that only about 1/3 of the Solana transactions are actual useful stuff for users and the rest are consensus messages. But my point is that these things are only "too slow compared to Visa" by a reasonably small factor now, like 6500 vs 900. Thanks.
0: https://usa.visa.com/dam/VCOM/global/about-visa/documents/ab...
But it's been proven to be technically feasible to replace OldFi with blockchain, at least in terms of TPS.
Please cite your sources.
edit for people finding this later: Visa and MasterCard do on the order of one billion transactions per day combined. As far as I can tell, Ethereum (proof of work as of today) does about a million per day and Avalance (one of the top 3 proof of stake networks according to Wikipedia) does about a million per day. That's literally one thousand times fewer transactions than the top two card networks. Three orders of magnitude.