That's a good point, but I'm not sure it changes anything. Signal will fall back to SMS today if the target number isn't using Signal. That's one of the big reasons I use it, I don't need a separate app for my contacts that use dumb-phones.
Given that RCS will also need to fall back to SMS for phones that don't support it, the only distinction I see here is whether it's easier to get all of my friends to move to a new phone that supports RCS, or if it's easier to get all of my friends to switch apps.
Either way, my friends that refuse to change will need to get SMS messages, as they already do today.
On the Matrix side of things, this is theoretically solved by making the protocol separate from the client. This allows any company or phone to have their own Matrix client, and allows users to swap out their clients if they have other preferences.
For existing clients that don't support the protocol, you use bridges. There is already a Matrix to iMessage bridge, although it's really not practical for most people to use because Apple locks down iMessage so hard. But the same question applies -- will it be easier to get Apple to switch to a brand-new protocol, or will it be easier to get Apple to support a bridge to their existing protocol?
Of course, in the worst case a Matrix client could fall back to SMS for unsupported contacts, which (again) RCS is already going to need to do.
The problem with Matrix pretty much comes down to security. I keep hearing people say that they're going to turn on E2E encryption by default any day now, and as far as I know it still hasn't happened. My impression is the technology is still a little immature to rely on for security-critical tasks, although when the alternative is SMS, that may not matter.