I disagree that the bank models are broken. At face value, a more effective method is to pay the fee to the issue to get the updated card, if it exists. Nearly every issuer supports this nowadays. If you've ever wondered why your subscription magically updated to your new card after your old one expired, then this is precisely what they did. It's easy to do this with Stripe as well.

If customers care deeply about their subscription and you support adding multiple cards to ensure payments don't fail, then great. I'm not sure this makes sense as a product, though. It's very easy to implement this sort of dunning, roll over to next card, rinse-repeat strategy. The TCO of an additional vendor in this critical path is unlikely to be worth it, especially with the direction the merchant-of-record-as-a-service industry is headed.

In any case, one path I was insinuating is that these failure cases are good opportunities for customer engagement. Often times it can be as simple as the card-holding manager left a the company and no one realizes the sub is on their account. Depending on your response, this sort of event can make a huge difference to your funnel/lifecycle.

I wish you luck but I just don't think the market is there.

If, however, you also have an off-ramp to debt collecting then I think this is a viable product when targeting certain markets.