The more time goes on, and the cheaper actually running SSO becomes, the less this is true. Props to Github for allowing me to do SSO on my 1 man enterprise for $21/month.
Even if you have just 20 people, not having to manage separate sign in’s on all services is just so pleasant. Not pleasant enough to jump from $2400/year to $24k/year on all 10 of them though.
SSO is the only way to get 2FA working without the friction becoming prohibitive.
If SSO is a paid feature, only in some plans, you're selling an insecure product. You wouldn't make security patches exclusive to the enterprise plan, you shouldn't make 2FA/SSO exclusive either.
Extra security is a feature of enterprise plans precisely because enterprises are forced to buy them by compliance requirements (a good chunk of which is just security theater and blame shifting); no one else cares, people buy stuff, things mostly do not go wrong - a market balance is achieved.
I can see why this isn't ideal or desirable, but security maximalism also has a nasty habit of killing all utility of products and disempowering end-users, so I'm very much in the camp of trading security over other concerns.
Unfortunately, a need for SSO is about the only reliable way to gouge a large corporation. As a small fish you may like SSO, want SSO, you may even think you need SSO, but you really can get by without just fine. You're small - you can get around the requirements, or pivot, or whatever. A corporation is big and slow and can easily get themselves into a situation where not adding SSO will become a blocker for deals denominated in double-triple digit millions, but abandoning your product or the whole business segment will cost similar amount of money. In that situation, the vendor can have a field day milking the cash cow.