Ethereum Plans to Cut Its Absurd Energy Consumption by 99 Percent
The cryptocurrency is going on an energy diet to compete with more efficient blockchains
Illustration: Blood Bros.
Bitcoin soaks up most of the hype and the opprobrium heaped on cryptocurrencies, leaving its younger and smaller sibling Ethereum in the shadows. But Ethereum is anything but small. Its market capitalization was roughly US $10 billion at press time, and it has an equally whopping energy footprint.
Ethereum mining consumes a quarter to half of what Bitcoin mining does, but that still means that for most of 2018 it was using roughly as much electricity as Iceland. Indeed, the typical Ethereum transaction gobbles more power than an average U.S. household uses in a day.
“That’s just a huge waste of resources, even if you don’t believe that pollution and carbon dioxide are an issue. There are real consumers—real people—whose need for electricity is being displaced by this stuff,” says Vitalik Buterin, the 24-year-old Russian-Canadian computer scientist who invented Ethereum when he was just 18.
Buterin plans to finally start undoing his brainchild’s energy waste in 2019. This year Buterin, the Ethereum Foundation he cofounded, and the broader open-source movement advancing the cryptocurrency all plan to field-test a long-promised overhaul of Ethereum’s code. If these developers are right, by the end of 2019 Ethereum’s new code could complete transactions using just 1 percent of the energy consumed today.
Ether Evangelist: Vitalik Buterin, inventor of Ethereum, hopes to finally demonstrate the blockchain platform’s low-power format in 2019.Photo: Gordon Welters/laif/Redux
Ethereum’s attempted rebirth will be one of the year’s “most fascinating technologies to watch,” says Zaki Manian, who is advising the cryptocurrency upstart Cosmos. Manian says Ethereum’s development process means that multiple coders and organizations must collaborate in the open, converge on specifications, invent all of the technology to implement them, and make them work together seamlessly. “It is by far the most technically ambitious open community project that has ever been attempted,” says Manian.
Like Bitcoin, Ethereum relies on a blockchain, which is a digital ledger of transactions maintained by a community of users. (It’s called a blockchain because new transactions are bundled into “blocks” of data and written onto the end of a “chain” of existing blocks that describe all prior transactions.) However, Buterin designed Ethereum to do more than securely maintain a ledger without a central authority. His vision was for Ethereum to become a global computer—one that’s decentralized, accessible to all, and essentially immune to downtime, censorship, and fraud.
What gives the Ethereum blockchain such potential is its ability to store data, support decisions, and automate the distribution of value. It manages these tasks through smart contracts, programs written by users or developers in Ethereum’s custom coding language. Smart contracts have obvious business applications, but the long-term hope is that apps built from them will eventually make Ethereum the ultimate cloud- computing platform.
That lofty vision clashes with Ethereum’s current reality. While there are some multimillion-dollar apps running on it, even Buterin says he suspects that Ethereum is consuming more resources than it returns in societal benefits.
The problem is all that mining. Like most cryptocurrencies, Ethereum relies on a computational competition called proof of work (PoW) . In PoW, all participants race to cryptographically secure transactions and add them to the blockchain’s globally distributed ledger. It’s a winner-takes-all contest, rewarded with newly minted cryptocoins. So the more computational firepower you have, the better your chances to profit.
PoW mining is difficult by design. The idea is to prevent any one entity from controlling the blockchain. For example, if a bitcoin miner’s computer system had more than half of all the mining power on the network, that miner could perpetrate frauds, such as revising long- completed transactions. Bitcoin users would have little recourse because miners are anonymous.
In theory, PoW keeps mining a distributed affair. In practice, however, the development of application-specific ICs (ASICs) that accelerate mining, produced by a handful of chip fabs in China, has concentrated power over many cryptocurrencies.
Ethereum took the fight against concentrated power one step further by selecting a memory-intensive PoW algorithm for mining “ether,” as its value token is known. This ether-mining algorithm penalizes the use of ASICs.
What Ethereum’s PoW algorithm has not prevented, however, is explosive growth in the computing resources devoted to ether mining. The computational power directed at that task grew more than 25-fold in 2017, as the token’s value surged from $8 to $862 and mining firms built dedicated data centers full of general-purpose graphics processing units, which are well-suited to ether mining.
The resultant energy demand has created a backlash from environmentalists. Utilities and communities, meanwhile, see financial risk and opportunity costs if they cater to cryptocurrency miners that gobble up cheap electricity while creating few jobs. Serving miners may require utilities to make equipment upgrades, which could become superfluous if cryptocurrency prices crash and mining operations shut down.
Recent market dynamics support the utilities’ concerns. The value of ether peaked at $1,385 last January and then began a downward slide. In November it crashed below $120—low enough to erase miners’ profit margin and to prompt some to slow down or turn off mining rigs. According to a projection by the Digiconomist—a site created by Alex de Vries, a senior associate and blockchain specialist at PricewaterhouseCoopers —Ethereum’s miners likely cut their total energy consumption by more than half in less than 20 days.
No surprise then that some utilities, such as Montreal-based Hydro-Québec , are setting higher electricity rates for miners. Such pushback from utilities and their regulators may further erode the security of PoW-based cryptocurrencies. Restricted access to power and rising energy costs will hinder new miners from joining the game, accelerating the concentration of mining power. As it concentrates, the risk of collusion and fraud increases.
For Buterin, slashing energy use has been part of the vision from Ethereum’s beginning. Most of Ethereum’s other proponents agree. “It’s widely accepted in the Ethereum community that PoW uses far too much energy. For me it is the No. 1 priority,” says Ethereum contributor Paul Hauner, a cofounder of Australian cybersecurity and blockchain-development firm Sigma Prime.
Ethereum’s plan is to replace PoW with proof of stake (PoS)—an alternative mechanism for distributed consensus that was first applied to a cryptocurrency with the launch of Peercoin in 2012. Instead of millions of processors simultaneously processing the same transactions, PoS randomly picks one to do the job.
In PoS, the participants are called validators instead of miners, and the key is keeping them honest. PoS does this by requiring each validator to put up a stake—a pile of ether in Ethereum’s case—as collateral. A bigger stake earns a validator proportionately more chances at a turn, but it also means that a validator caught cheating has lots to lose.
Moving to PoS will cut the energy consumed per Ethereum transaction more than a hundredfold, according to Buterin: “The PoW part is the one that’s consuming these huge amounts of electricity. The blockchain transactions themselves are not super computationally intensive. It’s just verifying digital signatures. It’s not some kind of heavy 3D-matrix map or machine learning on gigabytes of data,” he says.
Slashing computational power and energy use is not just an ecological move. It also has a financial benefit, because it should reduce the rate at which fresh ether is issued to encourage validators—extra money that dilutes a currency’s value. “Because PoS validators aren’t expending all of this energy, we don’t have to reward them as much,” says Darren Langley, a senior blockchain developer with Rocket Pool, in Brisbane, Australia, which is developing an app that will assemble staking pools, paying interest to ether owners who join the pool.
Moving to PoS could also boost security. Under PoS, the location of each validator’s account is known and can be destroyed if that validator breaks the rules. Vlad Zamfir, Ethereum Foundation’s lead PoS developer, likens this to the Bitcoin community gaining the power to incinerate the data centers of a miner who abuses his power.
By 2015, the advantages of PoS had already convinced the Ethereum community to make the shift, and leaders such as Buterin had expected to do so in just a year or two. To make their intentions clear, Ethereum’s core developers reprogrammed their PoW code to create an exponential rise in mining difficulty. Known as the “Difficulty Bomb,” it began slowing the creation of new transaction blocks in late 2016 and was expected to bring ether mining to a grinding halt a few years thereafter.
This time bomb has, however, functioned more like an alarm clock with a snooze button. In October 2017, when mining time had already nearly doubled to 30 seconds, the Ethereum team reset the clock, delaying PoW’s doomsday by about 12 months. And they will likely hit snooze again shortly.
It’s not that Team Ethereum is sleeping in. In fact, Buterin says, Ethereum’s developers have already slain most of the theoretical dragons associated with PoS. But the process of turning those theoretical solutions into efficient software has been moving slower than expected.
What provides hope for 2019 is a radical new plan adopted by Ethereum’s leaders in June 2018. Before then, they had anticipated building PoS into the existing Ethereum blockchain. In June, they decided to make a clean break and to build an entirely new blockchain—one that operates solely via PoS.
The two-chain solution—dubbed Ethereum 2.0—makes a world of difference for Ethereum’s programmers because continuing on the original chain would have meant writing the machinery of PoS as a sophisticated set of smart contracts. Hauner, who is leading an effort called Lighthouse to build an Ethereum 2.0 software client, says Ethereum’s smart-contract language is a tough medium for writing complex code. “Writing smart contracts is a very constrained environment for computing. You can’t do complicated things on it,” he says.
Just a few months after the decision to shift to Ethereum 2.0, its PoS specifications had been sketched out and over half a dozen teams were already working on software implementations using a variety of programming languages. Hauner’s group at Sigma Prime is developing its Ethereum 2.0 client using Rust, for example. He expects this app and others to be running PoS on beta networks, or “testnets,” early in 2019.
Buterin says public testnets could be handling another Ethereum 2.0 innovation—chains with multiple branches to boost transaction throughput—by the end of 2019. But he warns that there could still be “unknown unknowns” lurking that could set their timeline back.
As a multibillion-dollar network, Ethereum obviously has a lot to lose if it launches glitchy or insecure technology. To play on Ethereum’s PoS chain, holders of ether will have to deposit a smart contract on the original Ethereum chain that irreversibly transfers ether to the new chain. Any missteps could jeopardize the entire ecosystem of developers and projects that use Ethereum’s smart contracts.
But Ethereum also has a lot to lose if it delays much longer. An array of well-capitalized projects—Cardano, Dfinity, Eosio, and Manian’s Cosmos, to name just a few—are hatching their own PoS-based blockchains. Like Ethereum, they seek to prove that high security and high efficiency are not at odds.
The first to unleash the potential for blockchain applications may well become the computing platform of the future. The others will probably wither away. “This environment is naturally quite predatory,” Manian says. “There will be a single platform that survives.”
This article appears in the January 2019 print issue as “Ethereum Will Cut Back Its Absurd Energy Use.”