Zero Emissions Grid Demonstration Project Follies: No Fraudulent Demonstration Projects Allowed!

Even as I regularly repeat my calls for a Zero Emissions Grid Demonstration Project, I’m ready for the next move in the back and forth. Suppose someone claims that a steady zero emissions electricity supply has been achieved? How can we determine and verify whether that is true? The facts can be sufficiently complex, and the incentives sufficiently perverse, that fraudulent claims are to be expected.

Consider the simple case of El Hierro Island. They set out in 2008 with the objective of building a wind/storage electricity system that would provide the island with zero-emissions electricity. To this day, the website of the wind/storage electricity company, Gorona del Viento, proclaims on its opening page “An island 100% renewable energy.” Proceed through the website, and you will find lots of happy talk about tons of carbon emissions saved, and about hours of 100% renewable generation. But if you are persistent, and finally get to the detailed charts of the latest statistics, you find that the percent of electricity from the wind/storage system for the most recent full year (2023) was only 35%. Because El Hierro is an island, it lacks the ability to cheat by sneaking in some electricity from gas or coal from a neighboring state or country and not counting it.

But now consider the case Switch Inc., which is one of the largest (maybe the very largest) companies that specialize in operating data centers. Like its colleagues in Big Tech, Switch is obsessed with the desire to show its virtue by claiming to have “emissions” as low as possible, preferably zero. As I discussed previously in posts here and here, the likes of Google, Microsoft, Meta, Apple and Amazon all have the same obsession, and they all put out annual “sustainability” reports that loudly proclaim their virtue in the headlines and introductions; but then, all of them ultimately admit in the fine print that their emissions are actually increasing with the voracious energy demands of data centers and AI.

Well, such honesty is not good enough for Switch. Go to their website here and you will find this unequivocal statement: “All Switch data centers have run on 100% renewable energy since 2016.”

Really? How have they accomplished that? Of course, you will not find sufficient detail in their own statements to check the veracity of their claim. However, Bill Ponton has done an excellent analysis at RealClearEnergy on August 6 definitively proving that their claim is fraudulent. The title is “Tech Titan’s Quest for Net Zero.” Although Switch has supposedly contracted for sufficient solar power and backup storage to supply the steady electricity requirements of its facilities, in fact basic math shows that they have not purchased nearly sufficient quantities of either to accomplish the job. Despite their claims to the contrary, they are thus sneaking undisclosed amounts of power from reliable hydrocarbon sources to keep their centers operational 24/7.

Checking into Switch’s claim of “100% renewable” energy for its data centers, Ponton focuses on a particular center (Citadel) outside Reno, Nevada. He goes to Switch’s 10K for 2021, where Switch discloses that it has contracted for 130MW of renewable (in this case, solar) power to run the facility. But is that enough? To figure that out, you would need to know what is the baseload power requirement of the facility, and also how much storage is available to turn the intermittent solar power into a continuous baseload supply. Switch omits that information from its 10K, but Ponton tracks it down in an article about the facility in Greentech Media for July 2020: the baseload power requirement of the facility is 30MW, and the available battery is a 60MW/240MWh Tesla Megapack.

So is 130 MW of solar arrays plus 240 MWh of storage sufficient to provide 30 MW of firm baseload power? Ponton goes through the calculations, and here is the conclusion:

For more than half of the year from September through March, solar generation is not enough to handle both daytime and nighttime demand of 720 MWh. Increasing battery storage from 240 MWh to 330 MWh will have some effect in reducing the system’s dependence on gas power backup, but above 330 MWh of battery capacity, the system is limited by its solar capacity. 

Switch has the option of increasing both its solar and battery capacity to reduce the percentage of gas generation required to back up the system. However, even with 200 MW of solar capacity and 420 MWh of battery capacity, gas will need to generate 1% of the total energy required to provide steady-state baseload of 30 MW both day and night.

So they would need to nearly double the storage capacity (240 MWh to 420 MWh) and multiply the solar generation capacity by more than 1.5 (130 MW to 200 MW), and still that would leave them needing to draw their supply 1% of the time from backup natural gas. Now 1% of the time may not seem like very much. But 1% of a year is 87 hours, which is close to 4 days. And for those four days, you need the whole 30 MW of gas power that you would need to run the data center the whole time with no solar power at all. You need to keep the gas plant fully maintained and ready to step in at all times. And you need to pay the gas plant’s full cost of capital even though it may be idle 99% of the time.

If I am reading Mr. Ponton’s piece correctly, even to get to the figures of 200 MW of solar arrays and 420 MWh of storage to provide 30 MW of baseload power, he is assuming (1) zero turnaround losses on stored energy, and (2) no such thing as a cloudy day reducing solar irradiance. I’ll let Mr. Ponton run the numbers on his spreadsheet, but I’ll bet that one good fully-overcast week in December or January could send the storage need from 420 MWh to a multiple of that.

Ponton provides this as a link to all of his calculations.

So the mighty Switch Inc. is exposed as no more honest about its assertions of zero emissions than our friend gkam. The moral is that we should accept no claim of achievement of zero emissions electricity from anyone who maintains a continuing connection to a grid with hydrocarbon generation on it. Otherwise, there is way too much potential for cheating.