With the rapid growth in demand for artificial intelligence (AI) and high-performance computing (HPC) processing power, more and more listed Bitcoin mining companies are beginning to explore shifting their data centers, power, and infrastructure towards AI computing. This trend has also led the market to repeatedly ask a question: Is AI changing or even reshaping the future of the Bitcoin mining industry?
In an interview aired on March 10th on the Blockspace Podcast, Liang Wang, Vice President of Canaan, a leading global Bitcoin mining machine manufacturer, shared his observations on this issue. He believes that Bitcoin as an asset is still very likely to experience new bull market cycles, but this does not necessarily mean that the Bitcoin mining industry itself will see prosperity similar to that of the past. A more important reason is not AI or HPC, but the fact that the economics of mining itself are gradually deteriorating over time.
The following was compiled by the Bitpush team based on the interview. It has been slightly edited for readability while尽量还原原意 (retaining the original meaning as much as possible).
"Bitcoin Will Have Bull Markets, But Whether the Mining Industry Will, I Really Don't Know"
Host:
Many listed mining companies are now shifting their computing resources towards AI. How do you view this change? Will AI change or even replace the Bitcoin mining industry?
Liang Wang:
I've actually been thinking about this question. First, I think we must embrace AI and AI HPC because it is indeed a huge change. It will change our way of life and also alter many professional structures; many jobs will be replaced by AI. This point, I think, is very obvious.
But if the question is whether AI will replace the entire Bitcoin mining industry, I personally don't think so. Bitcoin as an asset class still has value, and it has its own cyclical nature. So if you ask me if Bitcoin will have bull markets in the future, I think the answer is yes.
But if you ask another question: Will the Bitcoin mining industry itself have a bull market? Frankly, I don't know the answer.
Because what's taking attention away from this industry is not just AI HPC. A more important reason is that the economics of mining itself are gradually deteriorating over time.
"Entering the Mining Industry Now is Much More Difficult Than Five Years Ago"
Host:
Why do you think the economics are deteriorating?
Liang Wang:
Because the industry today is completely different from five years ago. Five years ago, if you could get a mining machine, you were very likely to make a lot of money. Many people in the industry really viewed mining machines as "money printers" back then. But that's not the case now; entering this industry has become very, very challenging.
Look now, the Bitcoin price has dropped to around $65,000 to $70,000, but the network's total hash rate hasn't decreased significantly, right? This in itself is telling. Logically, if the industry were really so clearly profitable, or conversely, if the industry were completely replaced by AI, you should see more obvious changes in the hash rate. But reality is not like that. The reality is, people are still keeping their machines running.
Why? First, because they need revenue. Even if they are not actually making a profit, as companies they still need operating income, need to maintain operations, need to keep their employees working. Second, it's because Bitcoin mining now plays an increasingly important role as a grid regulator. HPC runs continuously 24/7; you can't easily turn it off. Therefore, the grid actually needs flexible loads like Bitcoin miners that can be started and stopped flexibly to absorb peaks and troughs and help the entire power system expand. So many companies, even with thin profits or no profit at all, won't easily take their hash rate offline.
This is why I say the problem is not just AI stealing the spotlight from mining, but that it's becoming increasingly difficult for new miners to make money in this industry.
After 2028, Monetary Gains May No Longer Be the Key Driver of This Industry
Host:
What do you think the industry will look like in the next two or three years, or even after 2028?
Liang Wang:
Bitcoin has a mechanism that everyone knows about, but many haven't truly thought through the consequences: the halving that happens every four years. What does the halving mean? It means that if the Bitcoin price doesn't double, the economic收益 (revenue/benefits) the entire industry gets from block rewards is declining.
So everyone actually knows that the next halving will be around 2028. Then the question arises: If by that time, the Bitcoin price hasn't risen to $300,000, or even to a level high enough to support miners' earnings, what will the industry rely on to move forward? This is a question I keep thinking about.
My view is that Bitcoin mining will continue to exist; it will continue to be part of the energy landscape. But I don't think that after 2028, monetary gains will still be the core factor for driving this industry. I think it's more likely that the industry will continue to exist around several directions, such as grid balancing, waste heat recovery, home-use scenarios, or utilizing resources that traditional energy systems cannot use efficiently. But if you ask me if there will still be a prosperous cycle where everyone rushes in and miners generally make a lot of money, I'm really not sure.
Of course, I hope I'm wrong. I certainly hope Bitcoin rises to $500,000 per coin, and then everyone rushes back into the industry. But this part, no one can really predict; I don't have a crystal ball myself.
AI and Mining Are Not a Zero-Sum Relationship
Host:
A very popular说法 (narrative/view) in the market now is that AI will directly take away power resources from the Bitcoin mining industry. Do you agree?
Liang Wang:
I don't really agree with understanding this as a zero-sum game. Because AI HPC and Bitcoin miners are essentially not the same type of load. AI HPC must run continuously; it needs to be online 24 hours a day, and often you simply cannot turn it off. But Bitcoin miners are different. One of the biggest advantages of Bitcoin miners is that they can be turned off when needed and can also be started up quickly when there is excess electricity.
This is also why I've always felt that in places like Texas, Bitcoin miners are actually quite welcome by the grid and energy suppliers. Because they can help the grid digest imbalances. They can yield power resources when needed, and they can "consume" excess electricity. Battery energy storage is certainly also a solution, but from a cost perspective, it's much more expensive than Bitcoin mining. So I've actually always been relatively optimistic about the position of Bitcoin miners in the energy system. I think they are not only not being replaced by AI, but might反而 (instead) become more important in another way in the AI era.
So in my view, AI and Bitcoin mining can often coexist in parallel, and even complement each other, rather than one necessarily squeezing out the other.
North America Remains One of the Few Markets with Long-Term Predictability
Host:
Looking at regions, where is future mining growth most likely to happen?
Liang Wang:
We've certainly looked at many places and tried many things. For example, we had considerable experience in Kazakhstan in the past. But the problem is, many countries initially welcome miners because you help them utilize otherwise idle electricity. However, once local electricity becomes tight, or the political environment changes, mining quickly goes from "welcomed" to "targeted."
This is also why we value North America more, especially the US and Canada. It's not that there are no problems here, but overall it's more predictable. Doing business in the US, at least you know what the rules are, know the differences between states, and know that if you actually bring local employment, taxes, or help the grid, understanding and support at the local level is possible. But in many other countries, the uncertainty is too high. Sometimes, just a word, a policy, or even a change in attitude from a local power figure, and your entire business is gone.
For a listed company, this kind of long-term predictability is very important. You can't use shareholders' money to gamble on a market that you don't even know if you can still operate in ten years later.
"The Situation in China is Complex"
Host:
The outside world has always been concerned about the situation of mining in China. How do you view the situation of the remaining mining farms in China now?
Liang Wang:
The situation in China has always been complex, and I don't think this matter is a simple, completely top-down, uniformly orchestrated process. China is big, and considerations at different levels and in different regions vary. The national level is more concerned with financial stability, especially not wanting asset outflows achieved through mining or Bitcoin trading. This has always been a very important background. So after 2021, domestic Bitcoin mining and crypto trading are indeed officially prohibited; this hasn't changed.
But on the other hand, the local level has its own realities. For some localities, Bitcoin mining was helpful; it could create jobs, create tax revenue, and also utilize electricity that otherwise couldn't be used, especially during economic downturns. Many people were willing to see miners as a way to monetize power infrastructure. So why do you still see some activity existing today? Because there is indeed market demand.
But what I want to say is, we have never understood this situation as "China will welcome Bitcoin mining again." We don't see it that way. Nor do we think it's reasonable to bet our resources on self-mining in China. Because as a listed company, you cannot make long-term capital allocations based on guesses about whether policies might change. This is also why we have always focused on North America, rather than betting on whether some new crypto strategy will emerge in China.
Is AI Taking Away Mining Machine Production Capacity?
Host:
There's another question in the market: With the surge in demand for AI chips, will it become increasingly difficult for mining machine manufacturers to obtain chip产能 (production capacity)?
Liang Wang:
I don't think capacity will be completely taken by AI. In the short term, chip capacity will certainly have times of tightness and times of looseness, but over a five to ten year horizon, I think supply and demand will return to a relatively balanced state.
Foundries like TSMC and Samsung make long-term investment plans worth tens of billions or even上百亿 (over ten billion) dollars. They look at the next ten years, not next year or the year after. Mining chips are certainly part of their business, but never the entirety, and not necessarily the core part.
What's actually more critical is whether you have a foundation for long-term cooperation with the foundry, whether you have successful tape-out experience, and whether you can truly produce a competitive product with a process node. Because with each generation of advanced工艺 (process technology), R&D investment is astronomical, and the cost of trial and error is extremely high. New players, without this accumulation, will find it difficult to break in even if they have money.
This industry competes not on "I want to make the most advanced chip," but on whether you can make it and whether anyone will buy it once it's made. So I tend to see this as a matter of long-term technical cooperation, rather than simply understanding it as AI taking away mining machine capacity.
Twitter:https://twitter.com/BitpushNewsCN
Bitpush TG Discussion Group:https://t.me/BitPushCommunity
Bitpush TG Subscription: https://t.me/bitpush
