Bitmain's marquee Ethereum miner will soon be obsolete, putting pressure on the blockchain network's hashrate

Ethereum’s hashrate could see as much as a 20% to 30% drop in the next year as mining hardware giant Bitmain's key ETH miner becomes obsolete.

That's according to a new body of research on this front, according to Kristy-Leigh Minehan, mining researcher and co-creator of the ProgPOW mining algorithm proposal. In a conversation with The Block, Minehan contended that the projected hashrate drop would gradually take place as mining devices – including the popular Antminer E3 devices – begin to fail due to the increasing size of the directed acyclic graph (DAG) file.

Ethereum's Proof-of-Work (PoW) system is memory-hard. Each mining device needs to load the DAG file at the start of a mining epoch (30,000 blocks). The size of the DAG file rises after every mining epoch, typically by 72% per year. The current DAG size is 3.52 GB, according to data published by Investoon.com.

According to a recent blog post by altcoin mining pool 2Miners, Antminer E3 – a widely used Ethash algorithm ASIC (application-specific integrated circuit) device produced by Bitmain – will be rendered obsolete in April as a result of the device having difficulties processing the size of the DAG file.

2Miners started getting reports on Antminer E3 performance on the Ethereum Classic pool in late February, the blog post said. Given that Ethereum is ten epochs behind Ethereum Classic, Antminer E3 is expected to stop mining on April 8. 

The ongoing research by Minehan points to the possibility of Antiminer E3 accounting for a significant proportion of certain hashrate signals, which would exacerbate the hashrate decline after those devices stop working. 

How big of a drop?

Minehan is known for her contribution to the ProgPOW proposal, which aims to narrow the performance gap between ASICs and GPUs by as much as 90%. The proposed change has caused concerns over the possibility of contentious forks, and a number of Ethereum community members –including Uniswap founder Hayden Adams, Truggle Suite CEO Tim Coulter, and DARMA Capital co-founder Andrew Keys–recently published a petition against it. 

Lesser known is the Ethereum mining research that Minehan is currently conducting alongside Timothy Ireland and other researchers. They are building on Coin Metric's previous research on nonce analysis to figure out the proportion of Ethereum network hashrate that's attributable to Antminer E3. 

According to a draft paper shared with The Block, researchers found a thin band of signals that started in April 2017 and began to taper off around June 2018. The timing roughly corresponded to the Antminer E3's product announcement and shipping dates. From her personal experience of working with the supply chains at the time of Antminer E3's launch, she knew that Bitmain bought up to 125,000 units of memory chips. After gathering more data, Minehan revealed that around 7%-8% percent of those specific signals are tied to Antminer E3, although further research is needed to confirm the number.

Despite the need for further analysis, Minehan is, however, certain that increasing DAG size would have a significant effect on Ethereum's hashrate. In the long run, she said, there could be a 20% to 30% drop over time in Ethereum's hashrate as ASIC and GPU machines go offline – and not just the E3 models themselves.

"It's pretty fair to say that regardless of what happens, we're going to see a 20 to 30 percent drop," she said. "And it's not going to be all at once. It's gonna be a tapered scenario where machines slowly drop off. And this isn't just due to the E3. This is due to all the other ASICs and GPUs slowly falling down as well." 

Network implications

EthHub co-founder Eric Conner said in a blog post published on March 10 that he is, most of all, concerned that an immediate drop in hashrate would result in a short attack window where the difficulty of mining a block is low. 

However, Conner acknowledged that there's little incentive to conduct a 51% attack during this brief period of time because it doesn't make sense from a game-theoretic perspective to "burn down your investment (ETH) to attack a chain for an hour."

"This will be a very quick window and while it's known, I'm not sure what would motivate someone more to do it in that moment versus over the past 18 months when hash started to drop off the peak and rewards were cut," he said.

Minehan agreed that the possibility of a network attack is low and that she's more worried about opportunistic miners performing reorganization attacks on the chain.

"I think we're going beyond the days where people are going to attack the network," she said. “What we're going to see more and more is that people are just going to do opportunistic attacks to recover stolen and lost funds. That might be a service that miners eventually start selling.”

Besides, Minehan believes that the short-term increase in block time could be problematic for the DeFi ecosystem. “My concern right now is that the increase in block time is going to affect a lot of our DApps in the DeFi environment,” she said. 

Building a global supply chain

Fixing the Antiminer E3 issue is not as easy as simply having other GPU or ASIC devices come in and fill the gap, according to Minehan.

"The recent impending recession that's hitting the U.S. is going to impact how a lot of these companies will continue to operate," she said. "And we're also seeing miners getting pretty upset after their treatment over the last two years. Many have just left the Ethereum ecosystem for good." 

"I don't think we're going to see new participants entering into Ethereum at all," she continued. "What we're just going to see is existing miners maybe reinvest a little bit more in the network and that's it. And new ASICs will probably continue to be built and mine in secret."

The only real solution to the problem, she believes, is to strengthen the supply chain for Ethereum miners to make it as robust and distributed as Bitcoin miners.

"Vulnerabilities in the Ethash algorithm would need to be widely publicized so that every manufacturer has the ability to start building an Ethereum ASIC from the get-go," she said. 

Currently, ASICs are not well distributed outside China. Many devices, including Antminer E3, also feature shipping restrictions. According to Minehan, this would have to change.

"There has to be a way for ASICs to be distributed all over the world with no shipping restrictions," she said. "You would need global supply chains, warranty, good customer support, and probably the backing of a very seasoned distribution company."

Building up a robust supply chain, however, could be a prolonged process.

"This could easily take years of time investment and monetary investment,” she added. “Bitcoin has also had 10 years of history to build up their supply chain."

Ethereum is set to transition into a Proof-of-Stake system, with Ethereum 2.0's Phase 0 expected to launch in Q2 2020. But Minehan argued that there's still going to be a prolonged period of time before Proof-of-Work becomes extinct. Ethereum mining and the ETH1.0 chain in general, therefore, still deserve much attention as of now. 

"We're having this issue in Ethereum right now where most of the client developers have just said they don't want to support 1.0 and want to spend all of their time working on 2.0.," she said. "2.0. Phase 0 is not proof of stake fully. Proof of work is still going to be there...It's a long way out, so until that point, we have to keep supporting the 1.0 chain."