The problem is that its not a DC connection to the grid.
Not only do you have to match voltage on re-connection, but you have to match frequency AND phase.
Its "doable" but to do so reliably without a stable reference can be challenging. That's why the DC requirement is there. They want some form of reliable power to well tested electronics to match phase and frequency when the grid comes back up.
A heavily loaded AC line will drift in voltage and frequency stability (say when your AC kicks on and browns out your solar panels).
I've read that before, and I have a further question if you could indulge me
What if, when the grid goes out the inverter keeps powering the house (and disconnects from the grid).
Then when the grid comes back on, kill the output of inverter (which results in outage to the house), reconnect to grid, then power up the inverter again, which skirts all the sync issues.
The house will see a few second outage... but it will have been "on" for the entire grid outage, which I think is much preferable.
The overall problem is more or less solved, it just adds equipment to the installation, both to make the power usable and also to ensure that it is isolated from the grid.
Not only do you have to match voltage on re-connection, but you have to match frequency AND phase.
Its "doable" but to do so reliably without a stable reference can be challenging. That's why the DC requirement is there. They want some form of reliable power to well tested electronics to match phase and frequency when the grid comes back up.
A heavily loaded AC line will drift in voltage and frequency stability (say when your AC kicks on and browns out your solar panels).