Cryptocurrency mining isn't dead—it's just hiding in office buildings in Shanghai
TechFlow Selected TechFlow Selected
Cryptocurrency mining isn't dead—it's just hiding in office buildings in Shanghai
Mining never died; it just changed its clothes.
Navigating Web3 tides with focused insightsBy Liu Honglin
For many, cryptocurrency "mining" still evokes images from the Bitcoin era—chasing energy sources like nomads: using wind power in the deserts of Northwest China during winter, and hydropower near rivers in Southwest China during summer. Thousands of machines crammed into metal sheds in remote deserts or riverside facilities in Sichuan, roaring day and night, consuming electricity like a flood.
But reality today is different. What's increasingly common in the industry is a form of "lightweight mining": no reliance on hydropower, no retreat into remote mountains—just a few quiet devices running in office buildings within cities, without fan noise or the smell of burnt circuit boards, silently "computing" and steadily generating tokens.
Due to my work, Lawyer Honglin frequently interacts with Web3 project teams, developers, and investors in Shanghai and Shenzhen. Many close friends invite me to visit their offices and proudly show me racks of hardware, introducing them as their cryptocurrency mining operations.
Outside the room lies China’s most centralized financial hub, bustling with traffic. Inside, silent machines run imperceptibly, producing neither sound nor noticeable heat, yet they support decentralized finance and dreams.
This "lightweight mining" model is actually an organic evolution within the industry under regulatory pressure. On one hand, large-scale deployments are no longer sustainable due to policy risks. On the other, as many new projects abandon Bitcoin-style PoW mechanisms in favor of lower-power alternatives such as PoS, distributed storage, and edge computing, the physical footprint of mining has become "invisible."
From a compliance perspective, this represents a classic gray area—equipment is compliant, networks are legal, and running nodes isn't inherently illegal. Yet the reward mechanism and incentive structure clearly fall within the realm of cryptocurrency. You can’t fully deny it's mining, but you also can’t clearly label it as illegal. This creates a delicate space for survival: operating quietly in the gray zone, not too big, not too loud, but very much alive.
To truly understand this reality, we need to revisit China’s regulatory trajectory on "mining."
As early as May 2021, the State Council’s Financial Stability and Development Committee explicitly stated in a meeting: “Crack down on Bitcoin mining and trading.” This triggered a nationwide systematic campaign to eliminate mining operations. Traditional mining hubs such as Xinjiang, Inner Mongolia, and Sichuan led the way, issuing power restrictions and shutting down mining farms. By September that year, the National Development and Reform Commission officially classified “virtual currency mining activities” as “elimination category” in the Directory for Guiding Industrial Restructuring, cementing the policy direction.
The official rationale was that such activities were “high-energy-consuming, high-carbon-emitting, and low-contributing,” contradicting national industrial policies and the “dual carbon” goals. At the time, this characterization had some basis in reality. The PoW mechanism dominated by Bitcoin was indeed emblematic of high energy consumption and density, with electricity usage surpassing that of some medium-sized countries, often powered by “gray” energy sources.
However, as the industry evolved technologically, many crypto projects moved away from PoW algorithms, adopting instead PoS, DPoS, and distributed storage for network maintenance. These models require significantly less computational power, shifting deployment from “remote metal sheds” to “urban office buildings.” You could call it mining, but it barely consumes any electricity.
What makes it even more complex is how AI development and surging demand for computing power have turned infrastructure once associated with crypto into something now encouraged by policy. Edge computing, distributed storage, and general-purpose GPU nodes—technologies originally developed for blockchain applications—are now being adopted and repurposed by the AI industry. At the level of computing architecture, the boundary between the two is inherently blurred: running an AI training model versus a blockchain validation node may use the same servers, differing only in software and purpose.
This creates a very real problem: traditional regulatory detection methods—such as “excessive power consumption,” “specialized equipment,” or “concentrated deployment”—are now nearly obsolete. It's impossible to distinguish which project is legitimately engaged in AI computing, which is covertly mining tokens, and which is doing both. Reality has already erased the regulatory boundaries.
So what we're seeing isn't so much a "revival of mining," but rather that "it never died—it just changed clothes." You’ll find many Web3 projects that publicly promote AI collaboration or edge node orchestration, but in practice are still executing blockchain validation logic. Others operate under the banner of data security or encrypted computation while effectively building their own token issuance mechanisms.
For local governments, this situation is equally challenging. On one side is the central government’s explicit ban on "mining"; on the other, strong policy support for "computing infrastructure" and "AI large model training." When a project sits at the intersection of both, there’s no clear answer on whether it should be supported, how it should be regulated, or if it even violates rules.
This ambiguity further pushes many projects toward a "run if you can, hide if you must" approach, giving rise to a more hidden, hybrid, and flexible underground mining ecosystem. You can’t detect it easily, quantify it clearly—power comes from residential grids, spaces are regular offices, finances are compliant, entities hold licenses—yet they’re still computing tokens. Applying traditional regulatory logic here simply doesn’t keep up.
As a legal compliance practitioner in the Web3.0 industry, Lawyer Honglin’s personal view is that among China’s “three bans” on cryptocurrency (ICO, crypto exchanges, and crypto mining), if any will see relaxation in the future, it will likely be mining.
Not because the state’s stance has shifted, but because the "new miners" have outgrown the original definition. It’s hard to describe them as “high-consumption, low-contribution” anymore. On the contrary, they may well be the very “computing entrepreneurs” you aim to encourage—receiving tech park subsidies, competing in AI contests, formally registering companies, paying taxes, and issuing salaries—except their profits include not just RMB, but globally tradable, liquid tokens.
Besides, as AI and Web3 grow increasingly intertwined, many blockchain teams are actively involved in AI model pre-training, data labeling, or algorithm optimization, while many AI firms recognize that on-chain incentive mechanisms offer greater efficiency in “crowd-sourced computing” and “edge participation.” Trying to forcibly separate Web3 from computing power is becoming ever more impractical.
Of course, I’m not suggesting full regulatory liberalization. Rather, we must acknowledge that the industry’s form has genuinely changed, and governing tomorrow’s reality with yesterday’s standards is no longer viable. Especially in ambiguous domains like computing infrastructure and AI service capabilities, the solution may not be blanket prohibition, but rather establishing a “positive list + industry classification” system—clearly defining which activities belong to the data industry, which fall under financial regulation, and which can operate legally provided they register and report.
Otherwise, if we continue equating “mining” with illegality and backwardness, we risk missing out on part of the future.
Mining today is no longer just a compliance issue, nor merely an energy issue—it’s fundamentally about how we understand the evolution of infrastructure. From Bitcoin’s “computing power for blocks” to the AI era’s “computing power as resource,” what we’re witnessing is the growing role of underlying compute nodes as universal interfaces of the digital society. If the past decade was about “who mines tokens wins,” the next decade may well be about “who controls elastic computing power holds industrial leadership.”
In an era of intensifying global competition over computing power, if the domestic market fails to build a mechanism that respects underlying technological pathways while bringing mining and computing into regulatory visibility, we risk being left behind in the next wave of global infrastructure competition.
Rather than blocking, we should first understand its true nature; rather than hiding, we should bring it into a transparent regulatory framework. This would at least allow projects capable of operating in the open to proceed with fewer concerns and less incentive for gray-area practices.
This is precisely the new question that urgently needs discussion.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
