The Future of PoW Mining After The ETH 2.0 Merge: An Overview

Proof of Work (POW) has been the shape of cryptocurrencies since the beginning. It is the default method for providing the consensus mechanism in the white paper of Bitcoin (BTC). However, with the massive rise in Bitcoin price and market cap, the competition for hashpower has been fierce, and almost 0.55% of global energy produced has been expended on it. This has raised huge concerns from the public about its energy sustainability. Recently, the European Union voted against a version of Markets in Crypto Assets (MiCA), a regulation that could ban cryptocurrencies based on Proof of Work (PoW) within the EU. This vote provided temporary relief for the PoW industry. At the same time, Ethereum, which is the second most common cryptocurrency after Bitcoin, is switching its mining algorithm from PoW to Proof of Stake (PoS) sometime in 2022. What is the future of PoW?

Bitcoin will most likely continue to use PoW although it costs a huge amount of energy

In order to answer this question, we first need to review the big picture starting with Bitcoin, the earliest experiment in cryptocurrency. Currently the Bitcoin PoW mechanism consumes almost 0.55% of global energy production, according to an article here. It’s a significant number and more than a lot of smaller countries’ consumption levels.

The key benefit of PoW is its reliability and security. Bitcoin, with its PoW mechanism, has been the longest cryptocurrency experiment in existence and has been vigorously tested since its start. It has been proven to be the most safe and reliable blockchain, with the most fair consensus structure. The PoS method has had a much shorter test period and has yet to find the balance between the trilemma of scalability, decentralization, and security. Bitcoin currently has a market cap of over 800bn USD and has the tendency to serve as a currency tender; thus, security is key, and it can afford such high energy consumption.

It’s also important to mention that part of the energy Bitcoin mining utilizes is energy that is other industries don’t really use. In the western part of China, there is ample hydro power that is uneconomical to transmit and can provide efficient green energy for mining Bitcoin. At its peak, this energy powered more than 50% of global mining before the Chinese government banned such activity.

After ETH switches to PoS, the remaining tokens will have a much smaller market cap and energy cost

The second largest PoW token currently is ETH, with a 300bn+ USD market cap, but this will soon not be the case anymore. The third largest market cap PoW token is Dogecoin; however, it has a much smaller market cap of only 16bn USD. Even Dogecoin is planning to move from PoW to PoS according to this article, which means there will be one less PoW member down the road. A 2020 research paper stated that BTC and ETH mining used around 80% of the total PoW power consumption. The significance of the remaining tokens is relatively small and accounts for roughly 0.1% of global energy consumption. The smaller this number becomes, the more difficult it is to prove the value of hanging onto the PoW method, which focuses on security. The logic is simple: you might want to buy full-coverage insurance for your newly bought Porsche, but you are less likely to do so for a second-hand, $1,000, 10yr-old car.

Table 1 - Top 17 PoW Token Data Ranked by Market Cap

Within PoW, ASICs vs ASIC-resistant algorithms fare differently

Furthermore, there is also a difference between Application Specific Integrated Circuit (ASIC) mining and ASIC-resistant mining algorithms. In Satoshi’s paper, he expected people to use their own computers, specifically CPUs, to mine for Bitcoin. While the CPU is good at multitasking, it’s not as efficient as a Graphics Processing Unit (GPU) for repetitive and high parallel computing tasks, and gradually people began to use their GPUs to mine. Then came FPGA (Field Programmable Gate Arrays), which brought in programmable chips that can specifically target certain mining algorithms and thus can be more efficient. FPGA paved the way for ASICs, which are the most expensive to manufacture while also the most efficient and powerful tool for mining algorithms. Nowadays, almost all Bitcoin mining is carried out by ASICs.

ASICs are expensive and efficient, which also brings in the problem of concentration. They are perfectly suitable for industrial management - a large number of ASICs running at the same site provide the best operational efficiency and economy of scale. However, such action leads to centralization, which goes against the decentralized nature of crypto and leads to increased risk of a 51% attack. ASICs could bring about a massive hashrate swing, as a few dominant miners might switch their mining tokens to find the best return.

A few token projects have specifically designed their consensus algorithm to be ASIC-resistant; this means they cannot be mined efficiently with expensive ASICs, but rather they are designed for GPUs mining, which is typically more decentralized. Currently, the tokens that are ASICs minable are LTC, ETC, etc; the ASIC-resistant ones are RVN, FLUX, etc. ASIC-minable tokens typically have larger markets than ASIC-resistant ones, as shown by the list above.

ASICs mining predominantly is carried out at a large mining site and costs huge amounts of electricity at each location. This energy expenditure is quite noticeable, and if it comes from non-green sources, there can be concerns about how sustainable that energy supply is. In contrast, for GPU mining, many miners are also gamers; their computers are used as gaming consoles when they play PC games and as mining rigs when they are not playing games. This system is relatively decentralized in nature and is less likely to attract serious attention and be seen as a major problem.

GPU mining is beneficial toward global GPU producers

A significant portion of the GPUs produced are sold to miners. Currently the two major global GPU producers are Nvidia and AMD. Nvidia is estimated to produce almost 70%, while AMD makes 30%.

In addition to regular GPUs, Nvidia has produced a Crypto Mining Processor (CMP) that is specifically aimed at miners. The company is fully aware of its role in mining, as stated on one official blog post:“‘We designed GeForce GPUs for gamers, and gamers are clamoring for more. Yet NVIDIA GPUs are programmable. And users are constantly discovering new applications for them, from weather simulation and gene sequencing to deep learning and robotics. Mining cryptocurrency is one of them.” According to Nvidia's filings, revenue from its CMP for 2021 totaled $550 million; however, when this revenue is added to that from GPUs sold to miners, the number is significant. An RBC analyst estimated that Nvidia had already gained 2bn USD in revenue related to crypto in 2018 according to this article.

Given the significance of the crypto market, it might not be a stretch to expect a certain level of support from GPU producers. In a total bear market, it might make sense for Nvidia to loosen its Low Hash Rate (LHR) limit for the newest 30 series cards so that miners are more likely to hold onto their holdings instead of selling them on the secondary market. Strategically, those PoW projects and GPU manufactures are ecosystem partners, as long as PoW continues to coexist. The similar logic works for ASICs manufacturers on ASICs friendly projects.

Once ETH switches to PoS, current ASICs, GPUs, and GMPs for ETH will have to find a place to continue mining or turn to a secondhand market

As of March 2022, the ETH hashrate reached 969MH/s, which represents something of an elephant in the room within the cryptocurrency circle. ETH can be mined either with ASICs or with GPUs that have more than 4GB of memory. After ETH’s transition to PoS, most ASICs that have been mining ETH can still be used to mine ETC. Thus, we might face an issue: where will the almost 1,000 MH/s hashrate go?

I carried out the calculation through a spreadsheet; at the current price, the breakeven hashrate for the market will be only 178 MH/s, a tiny fraction of ETH hashpower. In order to reach a breakeven point that would cater all the ETH hashpower, the price of major alternative tokens would have to rise 6 fold on average. Even at that point, the miners would just be able to make enough to cover their electricity bills with no profit. Fair to say there is elevated risk of a much lower profitability for miners going forward.

Table 2 - Hashrate estimate post ETH 2.0

In conclusion, a mix of things are likely to happen: 1), a few alternative tokens might see a significant price bump to attract more hashpower and become more mainstream; 2), a large number of GPUs will have to be resold to the market and end up in the hands of gamers; 3), a few tokens will likely lower their cost and go into PoS mode, following ETH and Doge.

Nothing contained in this website should be construed as investment advice.