https://www.inverse.com/article/34239-how-many-solar-panels-...
It's not if it would be nice to free up that land, but rather whether or not countries would be willing to spend the many trillions of dollars it would take, when they always have lots of other financial priorities.
I feel like if it were as easy as a 10x10 mile plot of solar panels to power the entire US, some enterprising disruptor would have done it already. Think of the revenue from powering the entire US...
You've also got to consider transmission and the infrastructure required, probably storage for time-shifting the energy as well, since it's solar.
At $250 per square meter, 10,000 square miles of panels would come to $6.4 trillion.
I think we should be heading in this direction anyway. But it's not so small, it's not so cheap, it will take time. No surprise; the system we need to replace is not small or cheap, and it took time, too.
Solar is expanding rapidly, and it will do so more rapidly as it gets cheaper. And wind is advancing rapidly too.
And remember, the article is about fusion electricity not starting up for another 30 years. By then, as my original comment said, it will be mostly or all renewables, and so there will be no way to persuade countries to spend trillions on fusion.
You wouldn't actually put all these panels in one place, because of transmission losses, putting all your eggs in one basket, etc...
The point being, that you can power the US with a reasonable, practical amount of panels.
[1] http://innovativesolarfarms.com/solar-farm-cost-per-acre/ gives $500k per acre
Musk's point was that even a 10,000 square miles isn't all that big compared to the rest of the United States.