51% attacks don't make sense because you're hurting yourself 51% and hurting everyone else 49%.
And for this theoretical attack you'd probably need to sustain 90%+ hash power for a long time.
Your first hurdle will be producing enough ASICs to surpass current hashrate by nearly 10x. Solve that problem and you'll need a massive amount of energy and you'll have to set up huge mining facilities in various locations to prevent crippling local power grids.
This would take years to plan & execute. A lot of people would have to be involved. Good luck keeping it a secret. Network hash rate will continue to increase while you're building this infrastructure.
If by some miracle you've pulled this off, Bitcoin users will switch to a fork of Bitcoin that uses a different mining algorithm and your entire investment is now completely worthless.