In 2025, a rare phenomenon occurred in several large Bitcoin mining farms: instead of miners scrambling for electricity, AI companies were lining up to request power from the farms.
No one anticipated that the cutting-edge AI models from Silicon Valley would ultimately be hindered by a seemingly outdated industry—electricity.
Thus, when the U.S. Department of Homeland Security recently announced an investigation into Bitmain, the surface reason was national security, but everyone in the industry knew it was just a facade. What truly made the U.S. anxious was that the future of AI was standing on a power system it could not fully control.
For the past decade, the U.S. has been competing for chips. But by 2025, everyone suddenly realized that while you could buy GPUs and rent data centers, you might not necessarily be able to connect to electricity.
The number of applications for data centers in states like Texas, Virginia, and Georgia surged, with some even facing a three-year wait to connect to substations. Electricity prices skyrocketed due to AI demand, and public power grids were forced to limit power to avoid collapse.
At this point, the U.S. finally realized that electricity is not just a resource; it is power. Whoever controls the power infrastructure holds the dominance in AI.
Meanwhile, Bitcoin mining farms, which are viewed as a marginal industry by the mainstream, had already quietly laid out a high-density power supply network of tens of gigawatts across the land.
Bitmain was founded in 2013 by two post-80s Peking University graduates, Jihan Wu and Micree Zhan. In just ten years, this company almost single-handedly brought Bitcoin computing power from household computer CPUs into the depths of specialization and industrialization.
At its peak, Bitmain accounted for over 70% of the global Bitcoin computing power, with profits reportedly reaching $3-4 billion in 2017-2018, surpassing Nvidia's GPU business during the same period. Its Antminer series was once considered a "money printer" in the hearts of miners worldwide, propelling both Jihan Wu and Micree Zhan onto Forbes' China Rich List.
However, as of 2025, this company, once regarded as a "business card for China's hard tech going global," has been accused by the U.S. Department of Homeland Security of potentially having remote backdoors implanted in its mining machines. If the U.S. and China were to enter a state of extreme confrontation, these hundreds of thousands of mining machines scattered across the U.S. could be activated as "distributed grid bombs," capable of causing local or even nationwide power grid failures through instantaneous high power consumption or harmonic attacks.
This accusation reflects a belated recognition of the strategic value of computing power itself. Mining machines are no longer just tools for mining Bitcoin; they are closely tied to electricity, energy, and national security.
The U.S. initially believed it controlled the world's most complex power grid, but as the growth rate of AI pushed the grid to its limits, it realized that the true understanding of how to harness computing power did not lie with Silicon Valley giants, but with those engineers who set up tents, propped up containers, and worked overnight to debug mining machines in the wilderness a decade ago.
The mining farm system does not rely on government projects or congressional budgets; it is a survival technology that has been squeezed out of the high-pressure competition and profit-driven environment of the crypto mining industry.
They possess incredible attributes:
- A project that takes a traditional data center three years to complete can be finished in three months in mining.
- Power scheduling can be adjusted in minutes based on electricity prices.
- Cooling systems can maintain stability in extreme weather conditions of 50°C in the desert.
- Grid connection plans can switch at any time according to policy changes.
- Data center structures can keep up with the rapid pace of hardware iteration.
And all these capabilities were not invented for mining. They were developed to combat the most expensive energy with the least amount of money. This is also what is most needed in today's AI era.
Thus, the U.S. suddenly realized an absurd but real situation: the infrastructure that AI is desperately seeking was built by the mining industry a decade ago, and it was constructed in the areas where the U.S. has the least control. This left the U.S. with a sense of "humiliation" in its passive participation in the future.
Therefore, the Bitmain incident is merely a way for the U.S. to express its anxiety—not about the company, but about the source of technology.
When they realized that the power base for future AI might need to rely on this system from Asia to sustain it, that unease was almost instinctive.
For the past twenty years, the proudest technological narrative in the U.S. has been "cloud supremacy," with chips getting smaller and smaller, model parameters inflating, and data centers being replicated globally like building blocks.
Until 2024-2025, tech companies like OpenAI, xAI, and Anthropic suddenly discovered that the biggest bottleneck in training trillion or even hundred trillion parameter models was not money, talent, or data, but electricity. Everyone suddenly realized that the true focus of technological history was shifting from the "cloud" to the "foundation."
And the most awkward reality is that 38% of Bitcoin computing power in the U.S., 70% of the world's professional mining machine manufacturing capacity, and the most mature experience in ultra-high-density power access are all in the hands of Chinese companies.
The mining industry inadvertently built the foundation that AI craves most, while the U.S. has yet to lay its own. This is the true meaning of the "Red Sunset Action."
Thus, we witnessed an extremely ironic scene: Bitcoin, once ridiculed by Wall Street and Washington as an "air coin" and "power-hungry monster," has, a decade later, become the sharpest needle piercing through American technological hegemony. It bluntly declares that the future no longer belongs to the country that writes code the fastest, but to the country that can provide the most, the most stable, and the cheapest electricity to that code.
The Bitmain incident is just the first horn call; next, global energy competition will completely shift from the shadows to the spotlight. The competition between nations will move from "model parameter scale" to "dispatchable power scale," and the core competitiveness of tech companies will shift from "number of engineers" to "substation capacity."
When electricity becomes the new oil, mining machines are just the first named casualties; the real battlefield lies ahead.
Related: Reports show: As record computing power meets falling Bitcoin (BTC) prices, the mining economy tightens.
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
