There have been quite a number of companies that have tried to do something similar, but with emulation on a hardware level instead of software - ie. you get a card that physically emulates your other cards. Coin and Plastc, for example, both failed, whilst Fuze is still going. The problem here is that it’s only really possible to emulate mag stripe cards. EMV chips are designed to be impossible to clone, so to emulate one, you’d need cooperation from the issuing banks, which is unlikely to happen.
Curve’s model of re-charging purchases to another card as an online purchase doesn’t require any cooperation from banks, but I believe the reason nobody else has tried it is probably because there’s no obvious way to make money out of it. Curve only get the interchange fee on each transaction (max 0.2%), but have to pay more than that to charge your underlying card. So they actually lose more money the more you use your card. Whilst other fintechs start off losing money, they at least have costs that scale sub-linearly with the number of users - if they get enough customers, the per-customer cost becomes negligible. Curve’s costs will go up with every user they gain, no matter how big they get. I’m not surprised nobody else thinks it’s a good business idea.
This is probably why Curve are trying to make money from selling insurance. The more popular they become, the more they lose. They therefore need to upsell people onto profitable products to be viable.