Transportation accounts for 25% of our energy use and greenhouse gas emissions and half of some pollutants like NOx and CO. That’s why there’s a giant effort to electrify it and power it with renewable energy. Many trains and trolley buses also are electric, but electric robotaxis offer a remarkable potential for powering our movement from the sun, with significantly lower emissions than any other form.

Renewable energy, like solar and wind, are now the cheapest forms of electricity to build, and they are going to get even cheaper. Financially they are an easy win, but they come with one flaw--they deliver when it’s sunny and windy, not when we ask for power. One solution to that is to have a way to store the energy, and the quest for this is a major industry.

Electric vehicles don’t need storage-- they are storage. It’s inherent in any EV for regular roads that it charge up at one time and drive at a later time, usually using batteries. The “problem” of solar is solved by the “problem” of the EV. For many EV owners today, charging can be a headache. Most charge at home or work, but 20% have to drive somewhere else to do it, usually at much higher cost.

Self-driving cars, however, will charge themselves. When they need energy, and it’s cheap, they will drive to stations where they can ideally plug themselves in. (Initially they will be “full serve” EV stations with human plug jockeys, just like gasoline used to be. Later the cars will do the plugging themselves.)

The chart above shows three things. The green line shows the California grid at the solstice. The line shows what percentage of grid power is renewable from solar and wind. (Not shown is hydro, and nuclear which is also low emissions.) As you expect, it’s a big hump mid-day when the sun is shining, and low at night. In fact, you’ll notice that on this particular day, the value is over 100%--there was a surplus of solar. More on that later.

The red area shows the passenger volume on a California electric train line: BART. Again, as you expect, most passengers travel at the two rush hours. It shuts down at night. Unfortunately, we can’t avoid that rush hours happen in the early morning and late afternoon, when there is some solar power but much less of it. The rush hour is tapering off when the real sun gets going. In the afternoon, the reverse happens, but it’s worse--demand is highest on the grid (especially in summer) late in the day, and it’s less renewable and most expensive right during evening rush.

Because the train does not use storage, it has to buy the power that is generated when the demand is high. That power is not as renewable as it could be.

The fleet of electric robotaxis can charge whenever they have time off. Today, power is cheapest at night, and cars are idle, so this is the most popular time to charge all EVs. We need to put more EV charging at offices and commuter lots for the regular drivers so they can use the sun. The robocars, though, can drive to charging when it makes sense. When it’s greenest, when it’s cheapest.

A typical plan uses almost all the vehicles during the rush hours. They are driving passengers, not charging. As soon as the rush ends, you still need lots of vehicles but you can send 20% of the fleet off to chargers for 30-40 minutes to pick up a charge. Then they go back into service and a new cohort goes off duty to listen to Cheryl Crow sing “soak up the sun.” If you can, you try to charge as many vehicles as you can from 10:30am to 1:30pm local solar time. During that time, there’s a major surplus of solar power nobody else wants. That’s the 120% number you see on the chart. During this time, the price of power on the wholesale markets is known to go negative . That means they have power that has to go somewhere, and they pay large wholesale parties to take it off their hands. It’s better than free. When this happens, every robocar in town that can plug in will want to do so. Free, zero emissions energy. The only limit will be how many charging stalls that can be built. The cars will drive a few miles to find them, it’s worth it. (Ordinary power customers can’t get paid to take power, but a large charging network could gain that status.)

We’re going to have a lot of that solar surplus because we need to overbuild our solar. We will, because it’s so cheap to build. You overbuild because solar panels make twice as much energy in summer as winter. Build enough for December and you automatically have a ton of surplus in June. We also need surplus to fill the storage. Today they pump the water back up into the reservoirs at the hydro plants; there are many other approaches built built. You also overbuild because you want enough on cloudy days, that gives you a surplus on sunny days. This means surplus and the word will get out on robocar twitter that power is going free or cheap or negative price.

Some people ask, “as we move to electric cars, how will the grid survive it?” This is one answer. In many cases, the cheap solar doesn’t even go through the grid. Cars drive themselves to the solar farms and drink up. Because electric cars are cheap to run, energy is a bit under 10% of the cost of operating a robotaxi, but it’s still worth getting cheap.

Working out the numbers on the sample day above, I found that the train (if it used power perfectly efficiently to match passenger load) runs about 60% on renewable power. That sounds good until you note that the California grid that day was 58% renewable on its own. In the plan above, some of the cars still charge at night, but even so, they ended up getting 86% of their energy from solar and wind. Often when people talk about solar powered cars, it’s a gimmick, but these can really be cars powered almost exclusively by the sun.

Of course, it’s not just cars. Today’s electric transit mostly draws power directly from the grid, and that’s in many ways a good thing. Transit can switch in the future from comparatively inefficient trains and buses to smaller, self-driving vans which provide vastly better service because they are smaller, frequent and non-stop. Providing better service at lower prices get many more riders than today’s systems--and the use of batteries means they can do it mostly powered by the sun.

It should be noted there are some wrinkles in this. Building all these vehicles and batteries has its own cost in energy and materials and emissions come from that. Studies have shown that this cost is quite reasonable, and won back in a very short time over fossil vehicles. The line-powered vehicles (like electric trains and trolley buses) gain many things from not needing batteries.

A real train line isn’t nearly so efficient as pictured. They actually run trains all day, with more empty seats, though not as many as at peak time. So they do use energy at noon, and that also can come from the sun, though it’s partly wasted moving empty seats. Their main problem is they have no choice but to draw lots of energy at 8am and 6pm when there’s not a lot of sun. Trains could have batteries to solve that problem, and the grid is building lots of storage for the same reason, but this is a large extra cost for the trains. For the cars and vans, having batteries is already part of the package. Trains have to get their power from the grid, they can’t go drive to a solar farm to pick it up. (Though train, owning railway ROW can sometimes install their own power distribution.)

That big solar hump in the middle of the day is very wide in summer, and narrow in winter. In winter, there’s very little sun during evening rush hour, and not much more in the morning. In summer, there’s much more, but there’s also tons of cooling demand in the evening rush--summer evenings are the highest demand period for the current grid, and will be until more air conditioners with ice (frozen water, not internal combustion) energy storage are put online. (If storing energy for cooling, water is somewhat cheaper than lithium, and also catches fire less often. It’s another thing that’s going to solve our grid growth problems.)

This is the reason the above plan doesn’t charge many cars by 4pm. Even though the sun is still up, power demand is so high that any power you take from the grid then causes more “peaker” fossil fuel to be burned, and the price can be quintuple the off-peak price.

This approach does require building many charging stations, since you try to use them mostly while the sun shines. The cost of such equipment will drop significantly at scale. Depending on charging speed, you would need 1 for every 8-10 vehicles. If they are costly, you don’t get to use quite as much cheap sunshine, but it still works.

This also suggests an interesting value for battery swap. Battery swap is not a great idea for private cars (though it happens in China.) But it can make sense for fleets, especially automated fleets. With battery swap, your batteries charge at the swap station, and can be set to do that only on the cheapest, greenest, surplus electricity. No charging at night at all. The actual swap can happen any time the car isn’t working.

This approach will apply to private cars as well, even those that can’t self-drive. Those that can do limited self-driving, ie. driving to charging stations on back streets, will still seek to charge during the solar surplus to get cheap power. Even those that can’t self-drive will try to get charging (ideally from local solar panels) wherever they park during the day, which is usually offices, commuter lots and homes of those who don’t commute. The big difference is that private cars only average around 35 miles/day (10kWh) while robotaxis will do 200 or more. The private car is very happy with low-cost “Level 2” charging, while a robotaxi charging mid-day will want something faster. Robotaxis charging at night may go with Level 2 if it’s cheaper.

This also becomes more important as utilities continue their fight against net metering. Without net metering, it’s essential to charge while the sun is shining to get the full value of solar panels, rather than selling the power back to the utility for almost nothing (or actual nothing) during a solar surplus. It becomes essential to build this structure, and to find robocars willing to drive up and take the surplus power, rather than throw it away at the utility. So people will find a way to do it.