If you think about it, it's the only thing that makes sense. The inverters are looking at the exact same grid the Gateway is. When the grid fails, the inverter(s) will sense this just like the Gateway does, and they will go offline as they're designed to on a grid failure. But when the PW-supplied microgrid comes up, provided the PW hasn't shifted the frequency beyond what the inverter(s) consider acceptable, they'll do their 5-minue check for a stable grid, and then start producing again. That's what happened with my system during simulated grid outage testing (throwing the main breaker). And the inverter log display showed it detected the line voltage at very low levels (like 5V or less) as the cause for the shutdown.
So normal devices/appliances that aren't watching the grid looking for faults would hopefully ride-out the Gateway's islanding and start of the PW's microgrid unaffected (though obviously not all things do), but something that's intentionally monitoring the grid looking for faults will see the grid loss and take action, there's really nothing the Gateway could do to stop that, since it has to see the grid fail before it can actually do anything about it. Inverters are on a 'hair-trigger' to shut off by design, consider that they'll curtail and/or shut down at 60.5Hz, when even most UPS's consider 63Hz "OK", and normal devices will run (even if not so happily) off even higher frequencies.