Vaults aren’t just any application, they are an attempt to improve the security situation for users protecting what are possibly their lifetimes savings, or perhaps billions of dollars worth of assets on behalf of others.
If people have bitcoin, they are implicitly trusting the core network itself to validate things according to the rules. There is absolutely demand for trustless vaults, which would be native to the chain.
If one relies on an oracle or other external source, that would require a different line of trust to said external system for the protection of coins. In a sense that is already available with BitGo or Coinbase or Casa, using various implementation of shared multisig schemes.
So we have evidence of implementation and robust demand for trusted, external vaults! But until they are implemented and available we of course cannot directly observe usage for trustless ones, we can only infer.
If we can reasonably replace a widely used trusted option with a crustless one, I think we should. Usage might well end up being higher for the trustless one owing to the superior security guarantees .