AI is killing miners: "energy bank run" tears open a new cycle, what fate is the mining industry standing at the crossroads of?

AI and Crypto compete for the same power pool, and miners are forced to choose: continue mining, rent electricity to AI, or completely transform the computing power infrastructure? (Synopsis: Trump hits the “loyalty” rating table for American companies!) Failure to cooperate with the policy may result in the loss of federal contracts) (Background supplement: What is the probability of independent miners successfully producing blocks? Bitcoin mining fully decrypted) Miners are being forced to choose. On November 18, when the global financial market was gloomy, Crypto once again ushered in “Black Tuesday”, and BTC fell below the integer mark of $90,000 during the day, falling about 30% from the all-time high ($126,000) more than a month ago, entering a technical bear market. At the same time, recent discussions on BTC shutdown prices, mining company liquidation and cyclical reshuffle have also begun to emerge, but unlike in the past, the core assumption of this round of discussions is no longer the internal supply and demand of Crypto, but comes from a more macro new fact: AI and Crypto have stood in the same energy pool to compete for power. It can be said that in the past year, from policy subsidies to energy/land tilt to marginal income differences, almost all variables have forced mining companies to complete a major shift from “electricity consumers” to “power providers” - and when “electricity training models” are more certain and profitable than “electricity mining BTC”, capital is also rapidly completing new positions. This means that in the coming years, miners will have to make a real choice under new structural pressures: continue mining? Leasing power and plants to AI? Or simply completely transform into a general-purpose computing infrastructure? This is perhaps a far more valuable question than predicting the movement of the Crypto market. What is the end of AI acceleration “squeezing” Crypto AI? Electricity. As early as 2024, Musk hit the nail on the head when he pointed out that the bottleneck in the development of AI is not computing power or algorithms, but energy supply, especially high-quality, sustainable, and data center power that can carry large loads. Therefore, an “energy run” around electricity, land, and factories is really happening. If you look closely, it is easy to find that AI data centers and Crypto mining farms actually have a high degree of genetic similarity in engineering implementation, that is, both rely on dense array computing units (NVIDIA GPU/ASIC miners), both need large-scale scalable park land/infrastructure plants, both need high-density, stable and long-term lockable power supply, and need a huge cooling/heat rejection/redundancy system. In other words, “first there is electricity and factory buildings, then there are computing units and customers”, this sentence applies to both mining farms and AI data centers, the difference is only that one uses electricity to train models to produce AI capabilities, and the other uses electricity to calculate hashes to produce BTC. This explains why mining companies have become one of the protagonists of this AI arms race: Crypto mines have long been prepared for a large amount of power supply, land parks, plant infrastructure, compared to starting from zero, technology giants can directly transform existing mines into usable AI infrastructure in a few months at the earliest, providing them with ready-made solutions that can immediately meet demand. Of course, this is not the first time that Crypto has deeply intersected with the “era of big computing power”, the last time was the Ethereum mining boom that opened in 2017–2018 “pulled” Huida's graphics card sales (Huang Jenxun even got a commemorative tattoo on his left arm because Huida's stock price exceeded $100 for the first time). However, this time the direction is just the opposite, it is no longer Crypto driving the computing power cycle, but AI in turn crowding out the energy space that miners rely on to survive. According to Morgan Stanley's calculations, if a mining company converts a 100 MW mine into a “live enclosure” data center (i.e., provides space, power and cooling, but does not include chips and servers), and then leases it to customers for a long time, it can create an equity value of about $5.19-7.81/watt, which is much higher than the current trading level of many bitcoin mining stocks. More seriously, in addition to market forces, the weight of the policy side continues to tilt: the United States has regarded AI energy as a key fulcrum of the strategic game, and subsidies, tax incentives, land indicators and power planning around AI data centers obviously take precedence over crypto mining. In general, the situation facing mining companies now is that they are in a deadly “sandwich” sandwich: Above is the AI dimensionality reduction blow: technology giants hold thousands of trillions of dollars in cash, are willing to pay several times higher premiums than miners to grab power contracts and transformer capacity, and greater capital gains are eager to strongly promote the transformation of mining enterprises; Below is BTC's own deflationary mechanism: the halving cycle continues to compress the currency-based income, the difficulty of the whole network continues to rise, the output per unit of computing power continues to decline, and the selling pressure caused by superimposed price fluctuations further blocks the living space of small and medium-sized mines; In this environment of multiple squeezes by AI, whether mining companies can find a new survival path is evolving into a fundamental question of whether they can cross the cycle, and three completely different development paths have been derived: Dead mining: continue to mine BTC, extremely low electricity prices, improve the efficiency of mining machines, and fight for the remaining living space after halving and increasing difficulty; Be a “second landlord”: lease your own electricity, plant and cooling facilities to AI companies or computing power service providers, transform into “power intermediaries + computer room service providers”, and earn stable rent and service fees; Radical role transformation: from a single mining enterprise, directly evolved into a general-purpose computing power provider, providing long-term computing power and hosting services for AI, cloud computing and high-performance computing (HPC) data centers, and truly turning itself into a new type of “digital infrastructure company”; Next, the valuation and fate of U.S. stock mining companies will largely depend on which path they are heading for. Second, the new valuation logic of mining enterprises: do not look at EH/s, only focus on GW/MW As mentioned above, mining companies seem to be forced to face a “three choices” question: continue to mine, sell electricity to AI, or fully transform? But in fact, under the impact of the wave of infrastructure created by AI, if you want to survive, there is only one ultimate solution to this problem: no matter which path you choose, you must complete the role switch from “power consumer” to “power supplier” in the next few years, otherwise you will be forced out before the next cycle. The reason is very simple - the power gap will be rigid for the next three years. Morgan Stanley's model shows that between 2025 and 2028, the power demand of US data centers is expected to be 65GW, but the current grid can provide only 15GW of near-term access capacity, plus about 6GW of data centers under construction, there is still a huge power gap of about 45GW. In the context of the rapid rise of the AI energy consumption curve, whether you can master electricity directly determines life and death. Fred Thiel, CEO of MARA Holdings (MARA), put it bluntly: “By 2028, you will either be a power generator yourself, or you will be acquired by a power generator, or a mining company that is deeply bound to the power generator and relies solely on the grid for power supply has entered the countdown to death.” To put it bluntly, the value of mining enterprises in the future will no longer depend more on the scale of computing power (EH/s) or mining machine/ratio…