I disagree.

Create friendly environment for miners to attract more hardware to the territory you control. Create government agency that will be responsible for helping miners become more environmentally friendly, f.e. by using heat produced by ASICs to heat homes. Give extra incentives in form of tax deductions when miners use their hardware that way - it would require them to register their hardware, so you'd know not only location but also its parameters and owner. Next, for example "due to widespread abuse of the system", require miners to run your software to monitor mining activity. Now you've removed the problem with acquiring enough hardware for the attack.

It's even easier in case of ETH, because its mining hardware can be used for other purposes, f.e. AI. Therefore you could even justify expenses in that field as part of military spending. Better yet, you can build couple of secret factories with that purpose in mind, and even if the secret leaks out no one would expect preparations for an attack on crypto.

On the other hand, if you want to acquire enough stake, you are creating an upward pressure that will push the price up and attract more investors to the coin you want to attack, even if you do it in distributed manner and over long period of time. You won't be getting your tokens with (enough stake)*(current price).

The main question would be: what are you trying to achieve with the attack? The example of Steem showed that even in case of very favorable conditions (large stake already in the hands of attacker, more pools of stake in the hands of his friends) and clear goal (take over governance) - the case of successful attack - other users of the coin can react simply by forking out. This is probably to strongest defense against any kind of attack (and such defense is easier with PoS, because it would be easier to know which funds were used in that attack, so you can omit them during airdrop to fork, just like Hive did to Steem).