From Manhattan Contrarians
Francis Menton
Despite my oft-repeated call for zero-emission grid demonstration projects, I’m ready for the next step. Suppose someone claims to have achieved a stable supply of zero-emission electricity? How do we determine and verify if this is true? The facts may be complex enough and the incentives perverse enough that fraudulent claims are to be expected.
Consider the simple example of Hierro Island. They began building a wind/storage power system in 2008 to provide zero-emission power to the island. To this day, the website of wind/storage power company Gorona del Viento claims on its home page “An island with 100% renewable energy.” Continue browsing the site and you'll find lots of great discussions about huge carbon savings and the time it takes to generate electricity from 100% renewable energy. But if you persist and finally look at the detailed chart of the latest statistics, you will find that wind/storage systems accounted for only 35% of the electricity generated in the most recent full year (2023). Because Hierro is an island, it lacks the ability to cheat without counting electricity generated by secretly importing natural gas or coal from neighboring states or countries.
But now consider the case of Switch Inc., one of the largest (perhaps the largest) companies specializing in data center operations. Like its colleagues at the big tech companies, Switch is keen to demonstrate its virtues by claiming “emissions” are as low as possible (preferably zero). As I've discussed in previous posts here and here , companies like Google, Microsoft, Meta, Apple, and Amazon all share this obsession, and they all publish annual “sustainability” reports, both in headlines and in profile They loudly proclaim their virtues; but then all of these companies end up admitting in the fine print that their emissions are actually increasing as data centers and artificial intelligence put huge demands on energy.
Well, that honesty isn't enough for the Switch. Visit their website and you'll find the following clear statement: “Since 2016, all Switch data centers have been run on 100% renewable energy.”
Really? How do they do this? Of course, you won't find enough detail in their own statements to check the veracity of their claims. However, Bill Ponton did an excellent analysis on RealClearEnergy on August 6th, clearly proving that their claims are fraudulent. The title is “Tech giants' quest for net zero.” While Switch purportedly has enough solar and backup storage contracts in place to meet the steady power needs of its facilities, in fact, basic math shows they haven't purchased enough of either to do the job. Despite their claims to the contrary, they still steal unknown amounts of electricity from reliable hydrocarbon sources to keep their centers running 24/7.
When examining Switch's claim that its data centers use “100% renewable” energy, Ponton focused on one specific center (the Citadel) outside Reno, Nevada. He talked about Switch's 2021 10K, in which Switch revealed that it has contracts for 130MW of renewable (in this case, solar) power to run the facility. But is it enough? To figure this out, you need to know what your facility's baseload power requirements are and how much storage is available to convert intermittent solar generation into a continuous baseload supply. Switch omits this information from its 10K, but Ponton tracked it down in a July 2020 Greentech Media article about the facility: The facility has a base load power requirement of 30MW, with 60MW/available batteries 240MWh Tesla Megapack.
So is a 130 MW solar array plus 240 MWh of storage enough to provide 30 MW of fixed base load power? Ponton made calculations and concluded:
For more than half a year from September to March, solar power generation was insufficient to meet the 720 MWh of demand during the day and at night. Increasing the battery capacity from 240 MWh to 330 MWh will have some effect in reducing the system's reliance on natural gas backup power, but when the battery capacity exceeds 330 MWh, the system will be limited by solar capacity.
The Switch has the option to increase solar and battery capacity to reduce the percentage of gas generation required to support the system. However, even with 200 MW of solar capacity and 420 MWh of battery capacity, natural gas would need to generate 1% of the total energy to provide a steady-state base load of 30 MW during the day and night.
Therefore, they need to nearly double their storage capacity (240 MWh to 420 MWh) and increase their solar generation capacity by more than 1.5 times (130 MW to 200 MW), but this still leaves them with the need to pump power 1% of supply. Now, 1% of the time doesn’t seem like much. But 1% of a year is 87 hours, which is nearly 4 days. During these four days, you would need 30 megawatts of natural gas power to run the data center with no solar power at all. You need to have your natural gas plant fully maintained and ready to step in at a moment's notice. Even if a gas plant is idle 99% of the time, you pay the full capital cost of the gas plant.
If I read Mr. Ponton's article correctly, even to get the numbers for a 200 MW solar array and 420 MWh of storage to provide 30 MW of baseload power, he assumes (1) zero turnover losses from the stored energy, and (2) ) There are no cloudy days that reduce solar irradiance. I'll let Mr. Ponton crunch the numbers on his spreadsheet, but I'd be willing to bet that if the weather had been nice in December or January, the storage requirements could have increased several times from 420 MWh.
Ponton provided this as the link for all his calculations.
So the mighty Switch Inc. is no more honest about its zero-emissions claims than our friend gkam. The moral is that we should not accept claims to achieve zero-emission electricity from anyone with a constant connection to a hydrocarbon-generated grid. Otherwise, the possibility of cheating is high.
Relevant
