It has high upkeep costs, requiring you to have your LN node online to receive payments, and for your node to have access to the LN funds via a connection to the private keys, which is a security liability.

To monitor for fraudulent channel closes in order to be able to react in time to challenge them, it also requires you to have an always on-line node, or a trusted third party delegated to do that on your behalf. And if you want rapid/automated reaction to fraudulent channel closes, that node needs to have access to the private keys to your LN funds.

Then there are routing issues when the topography of the network changes with every transaction, and where the existence of routes is dependent on sufficient capital being locked up in channels.

There's a reason why there's 100 times more BTC locked up for use in Ethereum dApps than for use in the LN. Ethereum-based scalability solutions could, ironically, provide BTC with the scalability needed to gain mass-adoption as an instrument for peer-to-peer payments.