I remember when high speed internet was coming into being and there were a lot of pundits talking about how high speed bandwidth would be sold as a commodity on the NYSE. If someone needed say an hour of high bandwidth for a video conference, they could do what you're saying, buy an hour of high speed access. Of course, high speed internet eventually became so cheap and so readily available, those ideas faded pretty fast.
I might be remembering this wrong, but wasn't Enron doing what you're talking about?
When 1,000 devices jump on the grid the moment electricity hits $0.01/kWh, the demand spike will cause more generators to come back on line and increase the price back to $0.05/kWh -- thus causing the 1,000 devices to drop-off the grid.
Rinse, lather, repeat.
How do you compensate for the potential grid instability?
Assuming instability does actually become a problem though, that sounds like a very straightforward technical problem with many possible solutions. Just off the top of my head, the simplest market-based solution would likely be futures trading. If all these appliances reserve their electricity usage 10 seconds in advance, then sellers can know exactly how much demand there'll be and adjust their production accordingly, maybe even bidding on that capacity so they know in advance exactly how much they'll need to produce. I imagine the high-frequency trading industry probably has tons of experience with this sort of thing.
I envision a future where individual appliances (including EVs) can opt in to spot pricing for the electricity they consume (or produce). That would naturally incentivize charging during off-peak hours and discharging during peak hours, all without requiring any government incentives or coercion. It could also be useful for other major appliances which could benefit from the lower prices afforded by load shifting, such as hot water heaters.