Blockchain’s Energy Crisis

Remember the oil crises of 1973 and 1979? Though much of the industrialized West and Japan felt the shock, gasoline was in such short supply in the United States that the federal government cut the national speed limit to conserve it and drivers lined up for hours to fill their tanks at sky-high prices.

Want a quick overview?

Read “Threat and Opportunity: Blockchain’s Hunger for Power.”

The blow to industrial economies was profound, but it also encouraged positive changes in energy usage, from limits on lighted advertising signs to greater fuel cleanliness and efficiency requirements for cars.

Something similar may be happening today as people realize that technology is a power hog. Data centers and transmission networks account for about 2% of global energy consumption, according to the International Energy Agency (IEA). Technology companies have become more efficient at meeting the demand for computing power and connectivity while using relatively less electricity. But as demand continues to rise, especially from the use of mobile networks, electricity usage and carbon emissions will also increase unless they are able to make additional improvements. By 2021, modest efficiency gains from data networks would still result in a 10% increase in electricity usage compared to 2018, generating around 9.9 million additional tons of carbon emissions. That’s the equivalent of driving 1.9 million passenger vehicles for a year.

Whether we are able to avoid a crisis like the energy shocks of the 1970s will depend on technology companies’ ability to increase supplies of electricity from renewable sources, deploy energy-efficient equipment, and use artificial intelligence (AI) to manage energy usage as the IEA recommends.

A shocking level of energy consumption

One now well-publicized example of technology’s power problem is blockchain. Blockchain’s use of intensive cryptographic computations to verify the existence and validity of transactions first attracted attention as the basis of cryptocurrencies like bitcoin. But now all kinds of industries are considering it as a solution to challenges of authentication, identity, and trust.

There’s one big problem, though: it’s the energy consumption equivalent of Grandma’s gas-guzzling 1973 Ford LTD.

What makes a blockchain work is the consensus of all the computers on its network that every transaction on the blockchain is true. The competition between computers to arrive at that consensus by solving cryptographic puzzles is known as mining. And mining uses powerful, purpose-built computer chips and software, plus a reliable Internet connection and air conditioning to keep CPUs cool as they churn away 24/7 at complex calculations. All of that consumes an enormous amount of electricity.

How much?

So much that if miners were a country, they would rank 41st in energy consumption, almost as much as Austria.

So much that mining generates as many carbon the city of Las Vegas or Hamburg.

So much that a single Bitcoin transaction consumes as much energy as 100,000 Visa transactions.

So much that cryptocurrency miners are setting up server farms in places with inexpensive electricity, like Iceland (which has ample geothermal power), rural China (which has underused hydroelectric power plants), and small-town America, putting a burden on their electrical infrastructure.

Indeed, one upstate New York town attracted so much mining activity that it briefly imposed a moratorium on new mining operations to stop them from literally using up the town’s electricity supply.

Some argue that estimates of blockchain’s power consumption are overblown, but even if that’s true, it’s still far from energy efficient. In fact, it poses an awkward conundrum: this technology was invented to make transactions more efficient is itself profoundly wasteful.

Current thinking about solutions

The problem may already be in the process of solving itself. First, there is the never-ending march forward in chip speed and efficiency, which delivers more compute power using less electrical power. Specialization helps, too. For example, Intel has filed a patent for a system-on-a-chip specifically optimized to use less space and energy for mining, joining similar initiatives from Samsung, Nvidia, and AMD as well as a proprietary chip from Bitmain, the world’s largest Bitcoin miner.

At the same time, the evolving blockchain industry is looking at less power-ravenous approaches to authenticating transactions. The original approach, proof of work, expends massive amounts of energy as miners compete to demonstrate that each transaction can be trusted.

One emerging alternative is proof of stake, which validates transactions based not on how much effort a participant in the blockchain puts in to solve the computational puzzle but on the percentage of their bitcoin ownership. Proof of stake would be more energy efficient than proof of work by eliminating the competition among miners.

Another alternative is proof of authority, which relies on a small but diverse number of known validators on a blockchain network who are rewarded for their work with transaction fees that aren’t related to the type or value of the transaction. Rules would be established to govern who becomes a validator and how they validate transactions.

Even if none of these alternatives pan out, it’s still possible that blockchain’s impact on carbon emissions will drastically decline over time. The mining computers that calculate and maintain blockchains can be located anywhere in the world, but because of their prodigious use of electricity, they’re increasingly set up where the cheapest power is available.

That could well drive energy production to be more widely distributed and decentralized, from hydroelectric plants on remote rivers to acres of solar panels in deserts. And that, in turn, could spur investment in new renewable energy sources and power storage for an abundance of cleaner, less expensive electricity.

On top of all this, blockchain could also make the power grid itself smarter, more efficient, and greener. Once we can generate low-carbon energy at scale, with devices enabled for the Internet of Things measuring production and demand in real time, we’ll be able to use blockchain to make sure power gets from where it’s produced to where it’s needed without human supervision. And that will further ensure that blockchain (and technology in general) becomes a high-efficiency engine for our economy instead of contributing to a global environmental crash and burn.