@seoincorporation, @PrivacyG and @OP, you three in my opinion are partly right and partly wrong.
The attack cost (cost to buy hardware/rent hashrate for an 51% attack) is higher than in Bitcoin's early days (OP is right here), and it is difficult to imagine a way to profit from an attack. However, it is maybe possible to imagine an attack where short sales play a major role. For example, imagine you short both Bitcoin, all wrapped Bitcoin chains/sidechains, most other cryptos (Ethereum, Solana ...) and also the stocks of mining pools and crypto exchanges. Today there may be too few coins to short, but in the future, if more mining pools and exchanges go public for example, then that may be different. So I think it's premature that the potential to attack Bitcoin is "completely lost".
This of course applies to all cryptocurrencies, not only Bitcoin. It is however still extremely difficult. We have seen 51% attacks only on small chains, the biggest having been ETC (Ethereum Classic) but in times when ETH still used PoW (so it was easy for ETH miners to attack ETC).
With @seoincorporation I also partly agree that price is not directly related to attack cost. But indirectly it is, because the difficulty and hashrate will grow in time if the miner income is higher as a consequence of a higher price. However, you also have to take into account block rewards (and its reductions by halvings) and transaction fees. (I think this is what @PrivacyG meant).
And with @PrivacyG, I agree with what you wrote about the hashrate/price relation. But this relation may change in the future if transaction fees become more important in the security puzzle. The period from the late 21. century on could be challenging for Bitcoin in this regard. Its security will be probably still high enough to fend off most attacks but lower than now. There are however possible mitigation strategies as I have just recently argumented in several threads, like a tail emission by merged-mining (mainly sidechains, but also other altchains).