On August 18th at 2pm UTC, people around the world are dialing in to the bi-weekly “core developers” Zoom conference call, which will be streamed live on YouTube to anyone who wants to watch it . None of the participants had their cameras turned on, and most appeared as black squares with names—including one called Vitalik, which hides Vitalik Buterin, the inventor of ethereum, behind it.
A handful of users adopted panda avatars, with cartoon faces shaking and smiling in front of their human counterparts. They chose this avatar thanks to Ethereum researcher Hsiao Wei Wang. He created a meme that shows two bears, one black and one white, dancing the “fusion dance” from the popular anime Dragon Ball Z. During the performance, the two bears fuse into a stronger individual with the dance. The panda – a combination of two bears – has since become a symbol of “merging”.
The merger is the crypto community’s name for the point in time when the Ethereum blockchain will transition from using “Proof of Work” as a consensus mechanism to using “Proof of Stake. They call it a merger because for nearly two years, a separate proof-of-stake blockchain called the Beacon Chain has been running alongside the original Ethereum chain for developers to test, improve and Test again. This Zoom meeting is all about getting developers to agree on when the two chains will merge. The date and time of the merger will depend on how much computing power is being used to maintain the blockchain, but the merger should happen around 1 a.m. UTC on September 15.
It’s not just technical fine-tuning. It’s a complete overhaul of a $200 billion software that’s been running for seven years, and if all goes according to plan, it will be implemented without downtime. People in the crypto space like to compare the process to changing an airplane’s engine mid-flight. Proof-of-work is energy-intensive and requires a lot of computing power, which causes blockchains like Ethereum and Bitcoin to consume as much energy as small countries. And proof of stake can reduce energy consumption by 99.9%. The impact of this on emissions will be overnight, with Ethereum expected to be shut down in the Netherlands after the merger (see chart). What’s more, if the merger is successful, it will demonstrate Ethereum’s ability to improve itself, opening the door for more radical change.
Cryptocurrencies need good news because the past year has been a tough one. A handful of dodgy deposit-taking VCs have collapsed; a cryptocurrency hedge fund has gone bust; stablecoins have been found to be unstable at all. The total market capitalization of cryptocurrencies has fallen to about $1 trillion, which is about $2 trillion lower than the same period last year. Improvements to Ethereum will not eliminate this disruption, however, by reducing its environmental impact and highlighting its potential for future improvements, it will show that the future of cryptocurrencies is brighter than people currently believe.
Buterin first released the idea of the Ethereum blockchain in 2014. Like Bitcoin, it is a large database of all transactions that have ever occurred in this cryptocurrency. But Buterin’s key insight is that blockchain can do much more than that — it can also track lines of code. This allows Ethereum to record the transfer of currency, as well as all assets and functions maintained in “smart contracts. A “smart contract” is a self-executing agreement that triggers a series of actions when certain conditions are met. By validating code on Ethereum, developers can build a large network of financial institutions, such as exchanges and lenders, with code on the Ethereum blockchain.
The blockchain is maintained by more than a dozen pieces of software called “clients” developed by core developers. Clients are built in a variety of programming languages, including Go, Rust, Java, and C#, and the software is run by “nodes” — computers that run client software to maintain the history of the Ethereum blockchain. All decisions about what to do and whether to implement upgrades are made by consensus among developers, ether holders, and those building applications on ethereum or listing real-world assets on the blockchain. Any plans and code are published live on GitHub, a repository for programmers. Core developers meet every two weeks to discuss potential upgrades. In theory, anyone can become a core developer by developing software.
The result is that developers are a mob. Some are employed by companies such as ConsenSys, a Brooklyn-based blockchain software company founded by Joe Lubin, one of the few people who helped create ethereum after Buterin published his white paper in 2014. Some of them are employed by the Ethereum Foundation, a non-profit organization founded in Zug, Switzerland in 2014. Ethereum is not a company and Buterin is influential and important as the founder of Ethereum, but not the CEO either. It’s open source — much like the free operating system Linux and the web browser Firefox — but provides an incentive for developers to participate in maintenance through the ability to buy shares of it with ether, but its governance is not truly decentralized Totally clear. When the project started in 2015, Buterin said he did a lot of research, thinking about what Ethereum should be, and doing a lot of coding to make it a reality. By 2020, he said he may have only conducted a third of his research, with little coding but mostly “advanced theory.” Over the past two years, he has said that even high-level theory “has slowly but surely slipped away from me”.
Implementing a change like a merge requires sufficient consensus from all parties involved. All major clients must agree to update software, enough nodes must update their software, and all real-world applications on the blockchain—like stablecoins backed by dollars in bank accounts—must accept New merged chains, which will keep their asset state. Watching this all happen in real time is surreal. It’s as if The Economist started live-streaming its editorial meetings and allowing subscribers to commission articles and choose covers.
Not all stakeholders are in favor of the merger. Miners have invested $5 billion in hardware to run the proof-of-work consensus mechanism. Around September 15th, the hardware will no longer be able to give them much in return. The way proof-of-work maintains blockchain security is by incentivizing hundreds of thousands of computers to solve a mathematical puzzle. The computer that finds the solution first alerts other miners, and if they confirm the result, update the blockchain and get paid. So it pays to have plenty of graphics cards in the lovely, new ethereum environment.
Proof-of-Stake is voted by cryptocurrency holders to decide whether to update the blockchain. Voting rights and share of rewards depend on how much ETH is staked. If stakers misbehave, such as making the wrong transaction, their stake is destroyed. So on September 15th, the advantage of having a lot of graphics cards will be gone. Instead, the advantage will be holding ether.
Miners may try to delay the merger by revolting. But most of these nodes seem to be cooperating with updates. According to ethernode, a website that tracks ethereum activity, about 75 percent have updated their software in preparation for the merger. Another option is to try to “fork” the blockchain, by still running the old software and hoping enough people do the same, the old version of the blockchain will live on. A hack in 2016 caused Ethereum to split into two chains: Ethereum (the dominant chain) and “Ethereum Classic” (the much smaller chain).
Justin Drake of the Ethereum Foundation said that for this split, “the world basically needs a miner to decide that they want to continue using proof-of-work,” meaning there will almost certainly be one such miner. The question is how many miners persevere and how many will give up. Chandler Guo, who backed the Ethereum Classic fork in 2016, is trying to organize miners around a proof-of-work token called “ETHW.” “I forked Ethereum once, and I will fork it again!” he said. While there are reasons for miners to stick with the old ways, the economics of trying to fork the chain may not increase. Mining ETHW only makes sense when the value of the coin is high enough.
Institutions like stablecoin operator Circle have backed the new approach, rather than any fork. In an Aug. 9 statement, the company said it “intends to fully support Ethereum’s proof-of-stake chain after the merger.” Wallet operators and exchanges also support proof-of-stake chains.
These dynamics reveal the inherent balance of power in Ethereum. Developers can’t make updates that are generally hated because doing so would result in a messy fork; miners can’t refuse an update if everyone else supports it. Decisions made by institutions that run applications on the blockchain, such as Circle, can help resolve disputes between the two camps. This is very different from traditional technology platforms. Apple can roll out updates that both iPhone users and app developers don’t like, and there’s very little those two groups can do about it other than ditch the iPhone entirely. There is no such thing as a “forked” iPhone.
Ethereum’s way of reaching consensus is “a messy and ad hoc process,” Drake admitted. But if things go well, there are huge benefits. As Ethereum is mined globally, the impact of its energy needs being eliminated overnight will not be too great. Almost half of Ethereum’s nodes are in the US; about a tenth is in Germany. Other countries, such as Singapore, the UK and Finland, are less than 5%. But in some small countries where mining is very popular, such as Singapore, energy prices could fall.
The change also reduces the need for specialized mining hardware. Gaming graphics cards from chip maker Nvidia can also be used for mining. From May to July, its chip revenue halved from the previous three months, partly due to rumors of an impending merger. Prices for used graphics cards are plummeting on eBay.
Since the network will no longer require as much energy and hardware to maintain, the rewards for validating transactions can be reduced. “Under proof-of-work, the scarce resource that is provided in return is computing power. It’s very expensive because you have to pay for electricity and hardware,” Drake said. With proof of stake, the scarce resource is digital currency. “So the maintenance cost is essentially the opportunity cost of that money, maybe 3% or 4%.” So the amount of tokens Ethereum pays to stakeholders after the merger will only be the tokens that were paid to miners before the merger 10% of the quantity.
This change in the monetary system may be one of the reasons for the surge in the price of ether since the timing of the merger began to be determined in mid-July. Even as Bitcoin and other crypto tokens traded sideways, the currency gained nearly 50%. Ethereum proponents believe that a successful merger could pave the way for a “turnaround,” where Ethereum’s market cap will surpass Bitcoin’s for the first time. It is currently valued at about half the value of its rival cryptocurrencies, approaching its highest share since 2017.
Another big benefit is security. Currently, to take control of a Bitcoin or Ethereum blockchain, an attacker would need 51% of the total computing power required to mine the currency. Rough estimates put the cost of the move at $5 billion to $10 billion. To attack the blockchain’s proof-of-stake would require acquiring and holding half of all tokens, which would cost around $20 billion.
Some argue that these benefits will come at the cost of centralizing power, as large shareholders earn more returns under proof-of-stake, further accumulating their holdings. But Ben Edgington of blockchain software firm ConsenSys said that claim is false. Small stakes will earn less than large ones, but their proportion of total tokens in circulation will remain the same over time, meaning their relative power will not increase. In a proof-of-work model, building larger, more efficient mining rigs also brings returns to scale. But Edgington points out: “It’s impossible for people to build a competitive mining rig at home.”
Another risk is that the transition fails in some way, which could erode public support. Lubin, co-founder of Ethereum, took this in stride. “We’ve done a lot of testing, and I think the blockchain element is going to work really well,” he said. Edgington believes the only missing link in the transition is the wider community. Given the complexity of installing new components and the need to master new ways of working, some players may be lost. But he said it would only be a problem if the churn rate exceeds 40%, which is unlikely. However, applications running on the blockchain, such as exchanges, may encounter some problems. Major software update reveals various bugs in code that previously looked fine. Some important DeFi applications, like lending platform Aave, are preparing to suspend ether trading during the merger.
If all goes well, the merger will be a step toward a more useful technology. Many financial applications running on blockchains are highly efficient, in part because they automate the functions of the financial system. Smart contracts automatically match buyers/sellers or borrowers/lenders on exchanges. A paper by the International Monetary Fund found that the marginal cost of financial intermediation through DeFi applications is about one-third that of banks in rich countries and one-fifth that of banks in emerging markets. However, the speed and cost of using the Ethereum blockchain can hinder user efficiency. When the network is busy, the fees for recording transactions are called “gas fees” and can cost as much as $100 per transaction.
The post-merger upgrades are primarily aimed at increasing scale and efficiency. At an ethereum conference in Paris in July, Buterin joked that the blockchain’s path is first “merge,” then “surge,” “edge,” “purge,” and “splurge.” “Surge” here refers to “sharding,” the process of splitting a database into chunks to spread the load. This will allow the blockchain to process more transactions and will reduce the fees required to use it. “Today’s Ethereum can handle about 15-20 transactions per second. This Ethereum … will be able to process 100,000 transactions per second,” Buterin declared.
“Edge” will implement a new type of mathematical proofs called “Verkle trees” and enable “stateless clients”. This would mean that someone could run software to operate a node without having to store the entire “state” of the blockchain, which is a huge amount of data. “Cleaning” will delete old data on the blockchain history. “Splurge” is “all the other fun stuff,” which is probably anything Buterin and the crypto-heads like. A successful merger is the first step in implementing all these changes. It will demonstrate that scattered groups of people can do risky, controversial and important things. Time to see if they can do it.