Banks are systemic actors in the financial system. They have privileges granted only to them, requirements that only they have to meet. It is HARD to build and operate a bank, whether in the US or in another country.
From a "fragility" perspective- there are a lot of definitions of fragility, but when looking at attributes like- may fail at any time; may lose customer assets; may take on inappropriate risks given their systemic position; have negative systemic impact on larger financial system operations- whatever attributes you think of, the amount of energy that goes into oversight around those attributes is significantly larger for banks than any other "systemic" industry (real estate, health, whatever).
The number of health oversight agencies in the US is, like, 5. The number of financial oversight agencies is in the dozens, just at the federal level. (This is why in the US you sue for medical malpractice but not financial malpractice, and why there is government backed financial insurance of various kinds and much less health insurance. You are on your own in the health industry. In finance there are armies around).
Banks have like 10 different kinds of risk that they and their supervision are monitoring for. Yes, banks fail, but that's because people suck and are bad at things and the second law of thermodynamics and so forth. It isn't necessarily the same in other countries but banking, like many other industries, is increasingly global, and banking core systems are increasingly international.
Nubank is clearly doing a LOT of things right- I saw a presentation in person from them in February and it blew my mind the level of capability and abstraction they were working at. They are a solid entity with a strong growth path.
Cognitect, in contrast, is a tiny consulting firm. Maybe 50 employees? Technical consulting at scale is an incredibly hard business, particularly when your solution domain is limited to a particular toolchain. There was really no chance Cognitect was ever going to become Thoughtworks, or Deloitte, or anyone like that. That's an infinitely more fragile financial situation for Clojure than living under the roof of a business with a valuation in the billions.
The way to think about this is that all technologies have a sponsor. Clojure moves from basically being sponsored by a guy in his backyard to, like PHP (Etsy has a market cap of around $10B and employs the PHP leader). That's huge.
Beyond that, having what should become a more well known and reputable player in the financial industry backing it- you could well see A LOT MORE adoption from enterprises playing copycat. It's not going to happen right away, but it is a seminal step in COMPLETELY changing the enterprise adoption calculation.
However, everything is a tradeoff. If you aren't working with Datomic, or in finance- this move is a stronger statement that maybe Clojure doesn't have the strongest future for you. It isn't going to become a lot of other things that it could become. The finance industry has a fondness for languages with weird syntax, and those languages often stay in the finance industry niche. That's probably where Clojure stays. There are all sorts of wicked cool JVM things like LMAX and Azul that are also financial industry specific, and Clojure and Datomic get to become one of those. Very successful, very cool, but niche.
Anyway- hope that's helpful, happy to continue the conversation.
Cheers.