Such a flood attack or spam attack is extremely costly. The attacker has to continuously outbid every other participant not once, but with a sufficient transaction volume to occupy all available blockspace. Meanwhile, each regular participant can decide to wait them out or increase the fee on their one or few transactions if their need to transact is urgent enough.
Not only is the attacker paying for 4,000,000 WU of blockdata per block at the highest fee rate, but if they maintain this level of fees, the increased reward is incentivizing miners to increase the hash rate, further increasing the blockspace supply that needs to be occupied. The attack is therefore largely self-defeating when aiming to suppress 99% of the transactions.
It may be useful for suppressing transactions below a certain level by creating e.g. a large backlog at 20 sats/vB, which would force other participants to overbid this level to get a timely confirmation. Such a backlog at a specific fee rate would be less costly to maintain and would get processed more slowly, but even so would cost up to about 29 BTC per day.