Is it really cost-effective to resist ASIC mining machines?

Almost all cryptocurrencies based on the Proof of Work (PoW) blockchain consensus mechanism have to face the same opponent - ASIC miners. As a Chinese chip giant that monopolizes the manufacture of ASIC miners, Bitmain's computing power and huge dominance over P2P networks may be dangerous to the entire community. The emergence of ASIC miners makes the Proof of Work protocol vulnerable to review and rule changes by a single centralized institution, thereby undermining the checks and balances between various stakeholders.

In fact, the developers behind cryptocurrency networks such as Ethereum and Monero have long resisted ASIC miners, believing that by using a memory-hard consensus algorithm, they can combat ASIC miners of "centralized mining" and make this mining method unprofitable. Some cryptocurrencies hope to try to completely abandon the proof-of-work consensus mechanism to solve the "mining tyrant problem". They will choose to use proof-of-stake (PoS), delegated proof-of-stake (DPoS) or Threshold Relay, etc., but these consensus mechanisms have not been fully tested and may also hide other centralization vector factors (other vectors for centralization).

On the other hand, there are some other benefits to using ASIC mining machines on the network. First, this professional mining hardware equipment is very efficient, and it has higher hash power per unit of electricity and better security. Secondly, ASIC mining machines are more reliable than homemade GPU mining machines, which can help miners carry out more targeted, professional and large-scale mining operations. In addition, ASIC mining machines need to rely on specific algorithms, so they are more flexible than GPU mining machines, and they can also allow miners to mine according to specific cryptocurrencies to obtain more block incentives. If we ignore the "political factors" caused by different interest groups in the crypto community, from a purely technical point of view, ASIC mining machines are actually more efficient and simple to operate when processing blockchain network mining, and at the same time, they will make the cost of hacker attacks higher, thereby ensuring better network security.

However, some industry critics of ASIC mining machines believe that such mining machines have triggered unfair competition, especially those large chip manufacturers can use the scale effect to squeeze out or even eliminate competitors. But in theory, if the creators of the blockchain network believe that a fair competition environment should be provided to everyone, then these blockchain networks that resist ASIC mining machines do not actually need to carry out boycotts. Riccardo Spagni, a well-known system maintainer of Monero, said:

“Bitcoin and Litecoin are the only two cryptocurrencies that treat ASIC mining machines fairly, but this situation may not happen to other cryptocurrencies now. And, it is entirely possible that we will not see other blockchain networks that can compete with ASIC mining machines for at least the foreseeable future.”

However, if the blockchain network creators who resist ASIC miners are wrong, for example, if mining equipment that can compete with ASIC miners appears on the market, does it mean that the mining market will become healthier and healthier, thus "forcing" cryptocurrencies such as Ethereum and Monero to eventually embrace ASIC miners? On the other hand, even if those who resist ASIC miners are right, and they also prevent the chip industry from being manipulated by boycotting, does it mean that Bitmain may be doomed, and those cryptocurrency networks that resist ASIC miners will eventually "unify the world", but does this mean that it violates the original intention of decentralization of cryptocurrency and blockchain industry? At this point, we can't help but ask, what are the right ways to solve these problems?

Next, this article will focus on the following six questions. We can also see from the answers to these questions whether it is really worthwhile to resist ASIC mining machines?

1. Why build a network that resists ASIC mining machines?

2. How is the work of resisting ASIC mining machines carried out?

3. Is the emergence of ASIC mining machines really inevitable for proof-of-work cryptocurrency systems?

4. What are the risks of choosing a network hard fork to resist ASIC mining machines?

5. Is it possible to create a fair commercial environment for ASIC mining?

6. What conclusions can we draw about the future of ASIC mining?

1. Why build a network that resists ASIC mining machines?

The core attribute of Bitcoin is the proof-of-work consensus mechanism, which allows computer networks around the world to reach a consensus based on a shared history. In order to prevent this consensus from being controlled by a single entity, the Bitcoin network uses mining to allow everyone on the network to jointly maintain network security and achieve network decentralization, which can theoretically resist the emergence of problems such as collusion.

As stated in the Bitcoin white paper:

“As long as the majority of CPU power is controlled by nodes, and those nodes are not cooperating to attack the network, they will generate the longest chain and outpace attackers.”

In the Bitcoin white paper, Satoshi Nakamoto described the proof-of-work algorithm as "one-CPU-one-vote", and the longest chain represents the majority of the network's decisions. In the event of a 51% attack, the attacker can take over the entire network, but if the cryptocurrency network is governed by many nodes (CPUs) running around the world, it is difficult to achieve this level of collaboration, because tens of millions of people around the world simply cannot communicate with each other (collude) and launch attacks on the network at the same time. Indeed, Satoshi Nakamoto built this important distributed network decentralized design concept, making Bitcoin seem difficult to be attacked by 51%.

However, driven by profit, Bitcoin mining has become a "horse race", and everyone wants to own a sweat-blooded horse - so GPU mining machines appeared in 2010, and ASIC mining machines appeared in 2013. This kind of hardware equipment equipped with a dedicated integrated circuit has a significantly higher mining efficiency than GPU and CPU mining machines. When ASIC mining machines were released, cryptocurrency mining became more centralized, and miners began to build data centers in areas with cheaper electricity prices. At this time, the decentralized concept originally designed by Satoshi Nakamoto to prevent 51% attacks began to be challenged, because when the computing power is concentrated in a few data centers or mining pools, there will be no "tens of millions" of independent miners in the network, and communication between "mining tyrants" will become easier.

With the emergence of ASIC miners, some bad results have also occurred in the cryptocurrency community. For example, people no longer want to use CPUs or GPUs to mine, and Bitcoin mining is no longer simply pursuing decentralization and equality, but has become more and more "centralized." In fact, if you want to participate in Bitcoin mining, you need to invest at least millions of dollars. In other words, only large mining companies have this level of resources to create market-competitive ASIC miners. What's more frightening is that these companies also firmly control the supply side of ASIC mining hardware sold to consumers. Compared with GPU- or CPU-based miners, it is much more difficult to create and use ASIC miners. After all, GPUs and CPUs can be easily purchased at local electronics retailers and assembled and run at home.

ASIC mining machines have led to mining pools sometimes controlling 51% of the computing power of the entire cryptocurrency network. Due to this centralized mining method, many newly established cryptocurrency projects have chosen the "ASIC-resistant" proof-of-work algorithm, as shown in the following table:

2. How is the work of resisting ASIC mining machines carried out?

It should be noted that resisting ASIC mining machines does not force hardware manufacturers not to produce such specialized equipment, but refers to cryptocurrency projects trying to change the mining algorithm so that ASIC mining machines are no longer profitable when mining. Currently, Litecoin, Ethereum, Monero and Dash have all confirmed (or rumored) that they will resist ASIC mining machines.

For example, Bitcoin uses the SHA-256 encryption algorithm, which is not an ASIC-resistant algorithm. When Bitcoin ASIC miners were introduced in 2013, their mining efficiency was several orders of magnitude higher (about a thousand times) than the GPU miners on the market. Using ASIC-resistant algorithms will narrow the performance gap between GPU-based miners and ASIC miners, which means that mining with GPU and CPU miners will still be profitable, although the profits may be much less than before. Taking Monero as an example, based on the new algorithm, Bitmain's CryptoNight X3 miner has "only" increased mining efficiency by 100 times compared to the current GPU chip-based mining equipment on the market.

In addition, when algorithms that resist ASIC mining machines are used, the cost of producing new chips will become very high, which means that few chip design companies can afford such expensive development costs from the beginning. The production of ASIC chips for different cryptocurrency protocols requires a large initial investment, which may be tens of millions of dollars, and the production cycle can be as long as three to six months, so it is difficult for companies that do not have sufficient funds to develop ASIC mining machines. In addition, the global shortage of silicon materials due to the growth of demand for artificial intelligence, the Internet of Things and mobile devices is also one of the reasons for the increase in costs.

Theoretically, the boycott of ASIC mining machines may end up with the following result: ASIC chips are extremely expensive, the noise generated by the operation of the equipment is huge, and only a few companies in the world are capable of producing them. On the other hand, we can find GPU chips in almost every household, and GPU hardware devices are easier to access and use for ordinary consumers. As a commercial hardware device, GPU is more widely used in daily life, and the production and purchase process is more decentralized. From this perspective, an equal proof-of-work system has actually been created.

At this stage, the main disagreement in the boycott of ASIC mining machines lies in the views of industry insiders on chip manufacturing. Even those who support the boycott of ASIC mining machines admit that commercial ASIC mining machines are almost impossible to achieve, and this kind of professional hardware equipment is easily affected by monopoly in every process link (development, production and distribution). At the same time, economies of scale and cheap electricity can allow a few companies to dominate mining forever. In other words, ASIC is fundamentally incompatible with the idea of ​​fair distributed mining, so it makes more sense to pursue GPU mining.

In the long run, resisting ASIC miners may not be a sustainable thing. Although there are many problems with deploying ASIC miners on cryptocurrency networks, even so, it is at least much better than trying to resist ASIC miners and ultimately failing. So what are the problems with deploying ASIC miners within cryptocurrency networks? ASIC miners actually combine miner incentives with specific projects. If a miner has many ASIC miners for the SHA-256 encryption algorithm, they can only choose to mine cryptocurrencies such as Bitcoin or Bitcoin Cash. However, if the network is successfully attacked, it will cause a price collapse. At this time, the ASIC miners themselves will become useless, and the value of the mined cryptocurrency will be greatly reduced.

In contrast, GPU miners are very flexible and can mine on many ASIC-resistant blockchains, so the attack vectors are larger. In layman's terms, even if a cryptocurrency blockchain is attacked, it will not make GPU miners useless because these miners can immediately switch to mining on other blockchains. In addition, there is another thing that is not recommended, although this is also an "advantage" of GPU miners to perform better: attacking some networks that are friendly to ASIC miners requires a lot of money for the attacker, while as long as there are enough GPU attackers, it is theoretically possible to perform a 51% attack at no cost.

3. Is the emergence of ASIC mining machines really inevitable for proof-of-work cryptocurrency systems?

In a successful and growing cryptocurrency network, the emergence of ASIC miners seems inevitable. Even if ASIC miners are not as efficient as GPU miners, it is possible to make it profitable to create specialized mining hardware and mine a certain cryptocurrency, such as SIA, a cryptocurrency project with a network value of $450 million. However, there may not be any ASIC miners running on the Vertcoin network, because the total market value of this cryptocurrency is "only" $100 million, and it may not be worth it for miners to invest in developing ASIC miners and mine on this low-market-cap cryptocurrency.

It is important to note that, contrary to what most people believe, it is difficult to detect the presence of ASIC mining machines on the blockchain network. "Smart" miners actually increase their mining output slowly over time, otherwise they will be easily discovered - when Bitcoin ASIC mining machines were launched in 2013, the network computing power suddenly increased sharply, which also attracted the attention of the entire community.

The same thing happened to Monero, which now uses the CryptoNight mining algorithm, which is effective against ASIC miners and ensures that CPU mining is profitable. But this was not the case at the end of 2017. At that time, the mining power of Monero increased by 400% from February to November 2017, reaching an all-time high of 1GHz/second. At that time, many people in the Monero community did not think that this situation was caused by ASIC miners, but thought it might be a "simple" price increase or the use of the Coinhive botnet.

However, on March 16, 2018, Bitmain announced the launch of the X3 ASIC Antminer specifically for the CryptoNight mining algorithm, which has a hashrate ten times higher than the most powerful GPU miner at the time. When Monero discovered this problem, it immediately decided to hard fork and change the proof-of-work algorithm on April 6, which also made Bitmain's X3 ASIC Antminer lose its original effectiveness for Monero. However, it is difficult to confirm how long Bitmain has been mining Monero using the X3 ASIC Antminer, and it is not known whether this type of mining machine caused a surge in the hashrate of the entire network.

When the X3 ASIC Antminer is officially shipped, it seems that it will only be suitable for some smaller, less profitable cryptocurrencies that still use the CryptoNight mining algorithm, such as Electroneum. It is actually difficult for miners to make a profit using expensive ASIC miners to mine these cryptocurrencies, but some people speculate that Bitmain may use the new ASIC miners to secretly mine for several months before releasing them to the market. Not only that, by selling these ASIC miners (the first batch sold for as much as $12,000), Bitmain may even make more profit than the block rewards obtained from mining.

Each mining company has a "price floor" to ensure that they always remain profitable. If a mining company's "price floor" is threatened, they will initiate a fork. A similar example happened to Vertcoin. In 2013, ASIC miners for this cryptocurrency appeared on the market. After that, developers began to switch to a new algorithm (Lyra2RE) and abandoned the previous algorithm. At present, there are no corresponding new ASIC miners on the Vertcoin network, but if its scale continues to expand, it may be targeted by ASIC miners.

In fact, for the proof-of-work cryptocurrency system, the emergence of ASIC mining machines is not inevitable. It mainly depends on whether the market value of the cryptocurrency is "worth it". If the market value of the proof-of-work cryptocurrency is small, mining companies chasing profits will not develop ASIC mining machines specifically. However, from the long-term development of the cryptocurrency market, ASIC mining machines cannot be completely eliminated, because as long as there is profit, there will be profit-seeking people flocking to it.

4. What are the risks of choosing a network hard fork to resist ASIC mining machines?

For those crypto communities that want to resist ASIC mining machines, how should they deal with the series of risks caused by this?

Generally speaking, the easiest way to resist ASIC miners is to change the proof-of-work algorithm through a hard fork. Because application-specific integrated circuits (ASICs) are only suitable for a specific algorithm, a cryptocurrency only needs to make a "little" change to make such mining equipment useless. This is different from GPU-based miners, which are more flexible and can be used to mine many cryptocurrencies with different algorithms, such as Monero, ZCash, Ethereum or Vertcoin.

However, changing the proof-of-work algorithm can resist ASIC miners once or twice, but this strategy is not sustainable in the long term. It's like a "cat and mouse game". The community not only needs to reach a consensus on constantly hard forking to change the proof of work, but also has good execution to do so. However, as the open source protocol continues to develop and become more and more widely used, this consensus will become increasingly difficult to achieve. Stakeholders in the cryptocurrency community may have realized that constantly hard forking to fight ASIC miners may be futile.

Not only that, there are also voices in the crypto community that believe that in a public chain network, the core development team should not rely on its own strong influence to repeatedly perform network hard forks. Bitcoin core developer Andrew Poelstra said:

“If the blockchain core developers have to change the proof-of-work algorithm every time an ASIC miner appears, then this approach actually makes no sense - because in a decentralized cryptocurrency, developers do not have this power; and in a centralized currency, proof-of-work is a completely unnecessary waste of power.”

In addition to the fact that it is increasingly difficult for the crypto community to reach a consensus that “hard forks are necessary because of ASIC mining machines”, there are at least four potential risks in implementing a hard fork to resist ASIC mining machines:

Risk 1: The network may introduce new vulnerabilities or bugs, either accidentally or maliciously.

Changing the algorithm every few months or every year may sound simple, but many things that arise from this are unknown and may even go wrong. Public chains should maximize resilience, which means that they need to be more conservative when making large changes to the protocol.

Risk 2: Hard forks will disperse the computing power on the network.

If ASIC miners are successfully removed from the network, the most obvious problem that follows is a significant drop in the total network computing power, which in turn causes the network to become unstable and in trouble for a period of time. There is no doubt that hard forks will disperse the network computing power, which will make network attacks easier. From this perspective, GPU and CPU miners may be more secure.

Risk 3: GPU mining will also be affected by vertical mining companies such as Bitmain, ultimately causing problems such as economies of scale and centralization.

If the developers of the cryptocurrency network insist on resisting ASIC miners and are determined to use GPU mining, then Bitmain is likely to enter the field of GPU development and mining, which will eventually lead to the same mining concentration problem. The most important reason why Bitmain has a huge advantage in Bitcoin mining is that they have huge funds and cheap electricity, which can be easily extended to GPU mining.

Risk 4: In order to adapt to some minor algorithm adjustment requirements, developers may need to build more flexible FPGA mining machines.

One problem with ASIC miners is that they only work for a specific mining algorithm. FPGA miners are actually miners that use field programmable gate array (FPGA) chips. It was one of the early miners, but it was not active for long and was quickly replaced by ASIC miners. FPGA miners are much slower than ASIC miners, but still faster than GPU miners. If the proof-of-work algorithm changes, three out of four ASIC miners may be devastated, while the fourth FPGA miner may adapt to the new algorithm. But this will produce a worse result, that is, the entire network mining will become more centralized.

5. Is it possible to create a fair commercial environment for ASIC mining?

If the resistance to ASIC mining machines is unsustainable, what is the most likely outcome? Does the existence of ASIC mining machines mean that centralized mining will not be completely eliminated? Not necessarily, because we may enter another situation: commercial ASIC mining.

"ASIC commercialization" refers to an imagined future market in which many different mining equipment manufacturers produce ASIC miners that compete with each other in terms of computing power and price. In such a conceived future vision, it means that no mining equipment manufacturer will have absolute "dominance" in terms of chip cost and computing power. Commercialization is actually a slow and gradual process, and the price of ASIC miners will continue to decrease over time. Dave Collins, a core developer of the cryptocurrency Decred (DCR), said:

“If the crypto community eventually accepts ASICs and intentionally makes them more efficient and cheaper, they will become commodities. Any single commodity will inevitably go through a centralization phase in a similar arms race process, but will eventually slowly move towards decentralization.”

Now, we can actually see that the competition in Bitcoin mining is getting more and more intense, and the degree of decentralization is getting higher and higher. This is mainly caused by the following three factors:

• Mining is more geographically distributed. Due to regulatory requirements, some mining companies are moving their operations to places such as Iceland, Canada, and the United States;

• The electricity cost for Chinese miners is no longer cheap. In the past, cheap electricity in mainland China allowed miners to have a great advantage in mining, making it impossible for miners from other countries to continue mining, but this situation no longer exists;

• Other chip manufacturers, such as Samsung and Intel, have also begun to enter the cryptocurrency mining market. Bitmain's huge profits have attracted these traditional chip manufacturing giants, who also want to get a piece of the pie.

However, the more important thing is: how long will it take for ASIC mining machines to be commercialized? How long will it take for ASIC mining machines to transform from a monopoly market to a truly competitive environment? In fact, this process takes much longer than you think. Bitcoin has only just started this process, and it has taken more than five years. Remember - it is just a start, and the subsequent work will take many, many years to complete. However, the longer it takes, the more vulnerable the protocol is to collusion and manipulation by a small number of miners.

Although it will take a long time, the commercialization of ASIC mining machines will eventually happen. As the market changes, some companies will succeed and some will exit the stage, just like other cyclical market laws. Let's take Bitmain as an example:

Before Bitmain was founded in late 2013, some companies in the cryptocurrency market had already begun producing ASIC mining machines, including Avalon, Butterfly Labs and Bitfury. When Bitmain entered this field, it actually felt that the market was relatively saturated.

However, an incident that followed completely changed the fate of Bitmain - the "Toumengou Incident". Due to the theft of the Mt.Gox cryptocurrency exchange, Bitcoin experienced a long bear market and the price trend became very weak. During the period of 2014-2016, many Bitcoin mining companies were forced to close, and the mining industry was also greatly consolidated. Thanks to its excellent products and scale, Bitmain successfully weathered the storm at that time. When the market shrank and profits became more meager, Bitmain achieved a curve overtaking and gained a large market share. When the market returned to a bull market and consumers began to consider buying mining hardware equipment again, the only thing available on the market was Bitmain's mining machines.

By 2017, the price of Bitcoin had risen by more than 1,000%, and a large number of competitors began to enter the chip manufacturing and mining market, and many new cryptocurrency mining businesses also sprang up. But what will happen when another long bear market comes? In fact, just like the last bear market, many mining companies will be forced to close due to reduced profit margins. In the current bear market we are experiencing, will new competitors emerge and surpass Bitmain like it did in the last bear market?

Bitcoin mining will eventually become decentralized and mining machines will become commodities, but this may be a very slow process. Samsung and Intel will compete with Bitmain, but it will also take some time. ASIC mining machines will be commoditized and widely used like GPU computing devices, and there will no longer be a single manufacturer, and mining will be distributed everywhere.

6. What conclusions can we draw about the future of ASIC mining?

Changing the proof-of-work algorithm is costly and an endless cat-and-mouse game. If the crypto community does not want to be trapped in this meaningless game, it needs to create a fair and sustainable environment for ASIC miner manufacturers and producers.

One way is to simply use ASIC-friendly algorithms and embrace ASIC miners with open arms, so that miners can use them in a more affordable and transparent way. As SIA developers have done, its network core development team is willing to take the initiative to promote and develop more friendly mining algorithms that support ASIC miners. But allowing ASIC mining to develop also means that cryptocurrency mining will become more centralized for a period of time when the market is immature.

Any solution is not easy to achieve, but after a series of evaluations, it is found that embracing ASIC mining machines may be the best way to go. Of course, "ASIC commercialization" is a very complex issue. Whether it can be realized in the end requires the joint exploration and efforts of various cryptocurrency project initiators, miners, and other stakeholders in the crypto ecosystem.