Mapping the chaos to find the signal in the noise — on July 20, Yangdian Tech (301012.SZ), a Chinese company known for smart lighting and energy-saving solutions, dropped a bombshell: it signed an 860 million yuan ($120 million) "computing power service" contract with an anonymous Client A. The contract runs for 60 months, representing a staggering 67.22% of its 2025 annual revenue. At first glance, this looks like a bold pivot into the hottest sector of the moment — AI and cloud computing. But dive into the fine print, and you'll find a familiar ghost: the undead corpse of Chinese crypto mining, wearing a new regulatory mask.
From the ashes of Terra, we learned to walk — but the ashes of China's 2021 mining ban still smolder. Yangdian's subsidiary, Sichuan Hanyang Smart Technology, is registered in Sichuan Province — historically the heart of China's hydropower-driven Bitcoin mining. The 924 Notice (2021) explicitly outlawed virtual currency mining activities. So how does a lighting company justify an 860 million yuan bet on "computing power"? The answer is regulatory semantics. By framing the deal as a general "computing power service" without specifying the end use, Yangdian skirts the legal definition of mining. The client stays anonymous, and the contract's deliverables remain opaque. This is not innovation; it's regulatory arbitrage dressed in corporate jargon.
Let's run the numbers. 860 million yuan divided by 60 months gives a monthly service fee of ~14.3 million yuan. Assuming this is for crypto mining (the most likely use case given the geography and industry context), what does that buy? At current Bitcoin prices ($65,000) and network difficulty, 14.3 million yuan (about $2 million) per month could support roughly 10-15 EH/s of hash rate if allocated to ASIC miners like the Antminer S21. That's a significant but not enormous operation — think 3,000 to 5,000 machines. But here's the kicker: electricity costs in Sichuan during the dry season run about 0.35 yuan/kWh, while the rainy season drops to 0.25 yuan. At 0.35 yuan, each S21 consumes 3.5 kW, costing about 8,800 yuan/month per machine. For 4,000 machines, that's 35 million yuan/month just in electricity — more than double the revenue the contract provides. Even if the contract covers electricity (unclear), the margins are razor-thin. The only way this works is if Client A provides the miners and pays for power separately, leaving Yangdian to collect a management fee. But then why the massive contract value?
Stories drive value, not just algorithms — and this story is being engineered for the A-share market. The narrative is irresistible: a small-cap industrial company reinvented as a "computing power provider", riding the AI wave. In China's stock market, such announcements trigger speculative frenzies. The stock will likely gap up 10-20% on the first day, attracting momentum traders and retail investors dreaming of a 10-bagger. But sophisticated investors should ask: what is the real asset here? Yangdian owns no IP, no proprietary chips, no data center expertise. Its balance sheet shows just 200 million yuan in cash and equivalents. To fulfill even a fraction of this contract, it needs to either raise debt or equity — diluting shareholders. And if the contract is just a hosting deal, Yangdian's revenue is capped at a fixed fee, while its liability is unlimited (equipment damage, regulatory fines).
When the crowd jumps, I look for the net — the network effects in this deal are nonexistent. Unlike Ethereum or Solana, where more users create value for all, Yangdian's computing power service is a closed loop. Client A is the sole buyer, making the entire business model a single point of failure. If Client A defaults after three months, Yangdian is left with idle machines and stranded power contracts. The anonymity of Client A is a massive red flag: in Chinese dealmaking, anonymous counterparties in billion-yuan contracts often hide related-party transactions or shadowy financial engineering. I've audited similar structures in DeFi lending protocols — the "anonymous whale" borrowing millions against flash-minted collateral. It never ends well.
Hunting for the next spark in the dry brush — let's consider the broader implications. If this contract succeeds (i.e., avoids regulatory shutdown and generates steady fees), it could trigger a wave of copycat announcements from other struggling manufacturers. Already, I've seen whispers on WeChat groups about two other companies scouting Sichuan for "data center conversions." This is a classic pattern: regulators crack down on overt mining, so the industry goes underground through service contracts. But the 924 Notice has teeth — the Chinese government can, at any time, deem any activity "related to virtual currency mining" as illegal. Yangdian's board is gambling that the political winds have shifted (with recent pro-AI rhetoric), but they're confusing AI compute with crypto compute. The former is encouraged by Beijing; the latter remains a perpetual gray zone.
Rebuilding the compass after the storm passes — here is my contrarian take: this contract is not a pivot to the future; it's a desperate lifeline from the past. Yangdian's core business has been shrinking — net profit dropped 40% year-over-year in 2024. The management is using a high-risk, one-client bet to mask structural decline. The A-share market will initially reward this story, but the risk-reward profile is abysmal for anyone buying above pre-announcement levels. The real alpha lies in shorting the stock after the initial pump, using the regulatory angle as a catalyst. Alternatively, if the mining narrative spreads to other stocks, one could trade a basket of "fake computing power" names — but that's a momentum game, not investing.
The map is not the territory, but the story is — for crypto natives, this case study reinforces a lesson from the past: institutional adoption often comes with baggage. The true signal here is not Yangdian's success but the regulatory arbitrage game that lets old mining capital rebrand. I expect to see more such contracts in Southeast Asia (Myanmar, Laos) where enforcement is weaker. As for Yangdian itself, follow the cash flows, not the headlines. If in six months they report zero operating income from this contract, the facade collapses. I'd rather own physical Bitcoin than a paper claim on a Sichuan mining farm with an anonymous landlord.
