This kind of construction has been discussed here:
The main issue with this construction is that clients need to come online and check the data availability. If they do not have data availability, the operator could withdraw funds from a state S_n and then roll back the chain to a state S_(n-k), where the operator has the withdrawn funds again in the side-chain. If the operator then also withdraws funds from S_(n-k), we get a “fractional reserve”.
Only if customers have data availability, they can stop the rollback. If they do not have data availability, they quickly need to request/enforce their deposit, in order to not be vulnerable for such an operator attack.
Still, even with this requirement of checking data availability regularly, it’s a great scaling solution.