Ethereum risks it all on going green

But sometime in early 2022, in an event that’s been rapturously termed “the Merge,” Ethereum will launch its most significant upgrade ever, promising to cut the network’s energy consumption by more than 99%, and positioning the blockchain, launched in 2015, as the big “green” choice for crypto users and developers. The product of more than five years of R&D, Ethereum 2.0 (or Eth2) has the explicit aim of making the fast-growing blockchain “more scalable, more secure, and more sustainable”—that is, if Ethereum can manage to pull it off without driving away its existing community. 

What’s behind the big shift

The most significant part of the update will be a fundamental change in the way the blockchain validates transactions across the decentralized network, a.k.a. its “consensus model.” Similar to Bitcoin, Ethereum currently uses what’s called a “proof of work” model. With Eth2, Ethereum will switch to a new “proof of stake” model, which will eliminate energy-wasting cryptocurrency mining altogether, replacing it with “staking.” 

Here’s the gist: When you pay someone with a debit card, or an app like Venmo, you’re trusting a central third party, like a bank, to validate the transaction—confirming that a payment request is legitimate, deducting the money properly from your account, and recording the transaction. Simple. In a decentralized payment system, such as a blockchain, the validation and recording of transactions are carried out by thousands of “nodes”—often called miners—across the whole network. In theory, none of these nodes—really, just people or groups with computers—“know” one another. So, how can the networks ensure that nodes behave honestly and don’t attempt to execute fraudulent transactions or otherwise bring down the network? 

One way is through proof of work. Currently, as transactions come across the Ethereum network, miners compete to aggregate, validate, and record them into a new cryptographically protected “block,” a permanent record of all transactions on the network made within a given time period. In return for doing this work, the miners get rewards, in Ether coin, plus additional transaction fees from users. The right to forge a new block is determined by what amounts to a high-tech lock-picking competition—a race to solve an intentionally difficult mathematical puzzle. 

The calculations require powerful computers and lots of energy to keep them running—and this expenditure of effort functions as “skin in the game,” presenting a barrier to entry that deters casual attacks on the blockchain and makes attempts to game the system unprofitable. Once a miner finds the solution for a new block, it is transmitted to all its peers across the network for validation. Each node performs a series of automated checks—and when everyone reaches consensus that the block is good, it is officially added to the chain. Miners that attempt to include a fraudulent block will have their block rejected, losing their reward, and the cost of electricity they’ve spent. 

The more computational power a miner has, the better its chances of “winning” blocks and racking up rewards. The rising price of Ether (up from $300 a year ago to about $2,200 as of late July) and other cryptocurrencies has set off an arms race, with miners amassing armies of specialized computers and often joining forces to improve their chances of mining block rewards (threatening blockchain technology’s utopian ideal of a decentralized network). In this low-margin game, energy cost is a key variable, so mining operations have tended to cluster in places where electricity is cheap, such as China’s Sichuan and Yunnan provinces, where hydroelectric power is abundant, or its Xinjiang province, where cheap electricity comes from coal-burning power plants. Before the Chinese government began its crackdown on crypto-mining operations this May, it was estimated that 65% to 75% of global Bitcoin mining occurred in China—and insiders that Fortune spoke with suggest that as few as six or seven firms handled most of the mining. 

Instead of having miners compete to solve puzzles for block-creating privileges (and rewards), Ethereum’s new model will rely on a network of “validators,” people or groups that put up a stake of at least 32 Ether coins (equivalent to nearly $72,000 as of late July). An algorithm will semi-randomly select who gets to create new blocks in the chain; validators who stake more than the minimum 32 Ether increase their odds of winning—sort of like buying more raffle tickets. New “candidate” blocks are validated by the rest of the network, similar to the old model in proof of work, but instead of getting a fixed reward for each block they mine, validators will receive an annual dividend based on the size of their stake. Validators discovered to be creating fraudulent blocks will be “slashed”—getting kicked off the network and having a portion of their stake “burned.” Validators that blow the whistle on bad actors get a reward. 

In contrast to the high-powered gaming computers or specialized rigs needed for mining, the technological hurdles for would-be validators is lower. An average laptop with a reliable, always-on Internet connection is all that’s really needed. The hope of Ethereum’s core development team is that this makes Ethereum more secure in the long run by allowing more people to participate in protecting the network. (Although there is nothing in Eth2 that prevents validators from joining together to create larger pools.)

A test version of Ethereum’s new proof-of-stake system, called the beacon chain, launched on Dec. 1, 2020, and has been running continuously with more than 16,384 active validators signed on to run simulations, minus real transaction data. The beacon chain will be merged with the current Ethereum blockchain, which holds Ethereum’s transaction history, to create Ethereum 2.0.  

Fixing Ethereum’s other problems

Ethereum won’t be the first crypto network to adopt a proof-of-stake consensus model. Many younger competitors already use it, including Solana (founded 2018), Ava Labs’ Avalanche platform; the Flow blockchain, created by NBA Top Shot developer Dapper Labs; and Cardano, launched in 2015 by Ethereum cofounder Charles Hoskinson. These upstarts are nipping at Ethereum’s heels, offering lower transaction fees and faster speeds to lure the next wave of blockchain-based app developers. 

As a result of its growing popularity among developers of so-called dapps (decentralized applications that run on a blockchain), the Ethereum network has become slow and congested. Right now, Ethereum can handle about 30 transactions per second at best; Visa, by contrast, does about 1,700 transactions per second and claims it can handle up to 24,000. Network congestion (Ethereum regularly handles over 1 million transactions per day) has driven up the fees that users need to pay miners to carry out their transactions. For those who want to carry out smaller blockchain transactions, the fees can exceed the amount of money actually being moved, which has driven some to seek out cheaper alternatives. 

“In the early days of blockchain, most of the activity was finance dapps doing more valuable transactions—tens of thousands, hundreds of thousands, millions of dollars,” says Patrick Barile, COO of DappRadar, a data acquisition and analysis company that tracks decentralized applications built across multiple blockchains. “This year, it’s more NFTs and emerging games that take advantage of the blockchain.” These consumer-driven apps demand fast transaction speeds and lower costs. As the first blockchain to offer so-called smart contracts, which allow you to run dapps, Ethereum had an early advantage with developers. “In 2015, that was real innovation,” says Barile. “But now developers who started with Ethereum are routing more transactions to other chains that embraced proof of stake from day one, which made them faster and cheaper.”

In late July at the Ethereum Community Conference in Paris, Ethereum cofounder Vitalik Buterin addressed these concerns, urging developers to grow the network beyond DeFi (decentralized finance) applications and emphasizing upcoming efforts to increase scalability and solve Ethereum’s high transaction fee and congestion issues. 

The Merge on its own won’t do that, though. The Ethereum 2.0 FAQ page states: “‘The Merge’ is limited in scope to upgrading Ethereum’s consensus mechanism. In practice, it will not have any effect on the current user experience of Ethereum today.” (Ethereum did not respond to requests for comment for this article.) However, a planned second phase of the Eth2 launch will directly address speed and fees, by introducing much-anticipated features and functionality. Those include “rollups,” which enable multiple transactions to be rolled into one transaction that is processed off the network but keeps the record of the transaction stored on Ethereum’s blockchain, and “shards,” a method of fragmenting Ethereum’s database for quicker retrieval. (Buterin has said that these and other improvements could eventually allow the network to handle 100,000 transactions per second.) 

These so-called Layer 2 initiatives, says CoinDesk analyst Christine Kim, “will help transition users and decentralized applications in the near term to a more flexible system that will be less decentralized and less secure, but will offer users the kind of low fees and fast transactions they need for their applications.” While other chains have an advantage in terms of speed and cost, says Kim, “Ethereum has amassed an ecosystem of dapp developers and user ecosystems that other chains don’t have.’ 

Grumblings in the community

Even before the Merge takes place, a controversial change to Ethereum’s fee policy called EIP-1559—being rolled out around Aug. 4 with the “London Fork” update—could lead to modest fee reductions, or at least more fee predictability for users of the blockchain. EIP-1559 will replace the miner-driven auction process that now determines transaction fees with a model that automatically generates minimum base fees for given transactions. 

Perhaps more significantly, though, it changes incentives for Ethereum miners: The transaction fees that once went directly to miners—save for optional “tips” to move up in line for processing—will now be “burned” by the network, taking them out of circulation and thus helping stem inflation of the currency. (Unlike Bitcoin, which is designed to stop issuing new coins once a 21 million coin cap is reached, Ethereum has no limit on the total number of Ether coins that can be minted.)  

Various estimates conclude that the loss of transaction fees could lower miner revenue by 20% to 35%. Even after the switch to proof of stake, with lower costs than mining, there is a risk that the changes to Ethereum’s incentive model could push disgruntled miners to leave the network, try to sabotage it, or start a competing chain. Indeed, an effort by some large-scale miners to organize a network disruption in protest against the fee changes in EIP-1559 prompted Ethereum’s developers to speed up the timeline date for the Merge, as a way to force would-be troublemakers off the network faster.

Most analysts expect that the shift to proof of stake will lift Ether prices, at least in the short term. That’s because the amount of Ether in circulation is expected to decline, as a result of large amounts getting locked up for staking, along with some Ether being removed from the system in the form of “burnt” transaction fees. CoinDesk modeling also suggests that new Ether will be created more slowly under the proof-of-stake system.

There’s no official date yet for the Merge—developers and analysts expect it will happen early next year. “But granted, right before they estimated early next year, they said the end of this year,” says Kim. “The technological risk is huge, so I wouldn’t be surprised if it’s delayed further. An upgrade of this scale is extremely risky. And hence there’s going to be a lot of pressure on developers to double- and triple-check their code. But beyond a shadow of a doubt, it will happen.” 

How things play out after—for users, developers, investors—is still unclear. Buterin has already been explaining that before any Layer 2 changes are implemented, there will need to be a round of cleanup work, introducing new code and protocols to make sure that the data transfer between the two networks functions efficiently. The much-needed Layer 2 will have to wait until that’s done. “Even after that,” Kim says, “there’s just so many promises—when it comes to energy consumption, when it comes to the issuance supply, when it comes to how this all impacts scalability—that will have to be realized. The markets aren’t going to price that all in until 2.0 is a reality. There are just so many empty promises in this space.”

A version of this article appears in the August/September 2021 issue of Fortune with the headline, “Ethereum’s big eco-bet.”