Does Bitcoin guzzle swimming pools of water? Depends what you measure, say critics of new study

A longtime critic of Bitcoin’s energy use found a new problem with the original cryptocurrency: It uses too much water.

“Bitcoin’s growing water footprint,” commentary published this week in online journal Cell Reports Sustainability, prompted a flurry of media responses.

But the industry cried foul, arguing the author — data scientist Alex de Vries, of the Netherlands’ central bank — had flubbed a past prediction about Bitcoin’s energy use and relied on a “debunked” metric to make hair-raising claims.

The debate has profound consequences for Bitcoin. Alarm over its energy consumption, and, in turn, its greenhouse gas emissions, has prompted calls from politicians to regulate the cryptocurrency or ban it altogether.

Swimming pools

According to de Vries, the total “water footprint” of Bitcoin in 2021 was 1,573.7 gigaliters. The Bitcoin network handled 96.7 million transactions that year, giving each transaction a water footprint of 16,279 litres.

That’s enough to fill a backyard swimming pool, according to headlines in the BBC, Euronews, and New Scientist.

The methodology was straightforward: Facilities that secure the Bitcoin network in exchange for newly issued tokens, a process called “mining,” used a certain amount of power in 2021. Based on the geographic distribution of those facilities, de Vries estimated the source of their energy, and the amount of water those sources used.

He was, in other words, attempting to measure Bitcoin’s “indirect” or “scope 2″ water use. “Direct,” or “scope 1,″ water use would involve the water used by the facility itself.

‘Debunked’

Critics say de Vries’ per-transaction methodology is flawed, and headlines that relied on it, misleading.

Bitcoin proponents such as Daniel Batten — an investor in Bitcoin mining operations powered by captured methane — point to the Cambridge Center for Alternative Finance, which states that “energy cost per transaction is not a meaningful metric in the context of [proof-of-work] blockchains.”

Proof-of-work is one of several technologies used to secure blockchains. In 2022, Ethereum switched to another mechanism, in part to reduce its energy consumption.

A single Bitcoin transaction “can include thousands of payments,” the Center for Alternative Finance said in a 2018 study. Additionally, “energy required for [proof-of-work] networks to function is independent from the number of processed transactions.”

“De Vries is neither a scientist, nor an energy or resource expert,” Batten told DL News. “He is the employee of a Central Bank, and has had his methodology of ‘per transaction’ debunked.”

De Vries defended his use of the per-transaction metric.

“New environmental disclosure requirements for crypto assets that are being introduced in Europe with the MiCA regulation will also specifically require efficiency metrics such as the electricity, carbon and water footprints per transaction,” he told DL News.

“Only a very limited fraction of Bitcoin transactions are payments, so taking transactions as payments tends to make Bitcoin look better,” he added. “Scaling solutions such [as] the Lightning network mentioned in the article have not had any meaningful impact in altering this.”

Direct, or indirect?

Batten also took issue with the notion that Bitcoin’s total water use was somehow problematic. Indirect water consumption is “not a meaningful metric,” he and others have argued.

“You could write an article saying that banking, AI, search engines, construction, media — any industry you wanted uses a lot of water using the methodology he uses,” Batten said.

“The actual direct water-usage of Bitcoin mining, which is what I would opine a piece of honest analysis would have focused on, is very low compared to other industries such as agriculture, pulp and paper, steam assisted oil recovery and extractive mining.”

But that’s just the standard across industries, according to Kaveh Madani, director of the United Nations University Institute for Water, Environment and Health.

Madani recently co-authored commentary on Bitcoin’s environmental footprint, which included an estimate of its worldwide water use in 2020 and 2021. That estimate detailed Bitcoin’s indirect water consumption — the water needed to generate the energy that Bitcoin used.

“That’s the correct way to do water impact calculations,” he told DL News. “To calculate the water footprint of the burger you eat … we go across the supply chain: So we start from the grass, and then the cow and the truck that brings the meat to the store.”

De Vries agreed.

“Assessing the water footprint in this way is no different than [how] we assess carbon footprints,” he said. “We have defined scope 1, 2 and 3 emissions to obtain a comprehensive overview on how entities impact the climate. … ‘It’s not fair’ is a lame excuse for avoiding taking any responsibility.”

‘Monitored and regulated’

Among other solutions to Bitcoin’s outsized water usage, de Vries suggested locating Bitcoin mining facilities in places with “favourable climate conditions to minimise direct water consumption”; “Immersing mining devices in a dielectric fluid” that can “provide cooling without relying on water”; and using power sources that don’t require fresh water, such as solar and wind.

Like de Vries, Madani says Bitcoiners should try to find ways to reduce the network’s energy and water use.

Still, he too had misgivings about de Vries’ commentary: For example, terms like “water footprint” and “water consumption” seem to be used interchangeably even though they have different, and specific meanings in environmental research — a red flag, according to Madani. (De Vries said his commentary “specifies the definitions of water consumption and water footprint in this article, essentially limiting what is considered as the water footprint and making the two terms interchangeable. Considering the full water footprint as understood by common definitions may result in a larger number.”)

And he wants researchers to find a way to communicate without antagonising the industry, lest it alienate its natural audience.

“We have to be extremely careful about the metrics that we present to the public, and the potential interpretation of those metrics,” he said.

“Research like this should deliver the message in a way that people do not think scientists are against technological disruptions, especially the positive ones. The goal here is to highlight some impacts that need to be monitored and regulated.”