Over the past 72 hours, Chengdu's municipal government dropped a policy bomb: a $36 billion (2600 billion yuan) AI+ action plan targeting 70% smart terminal penetration by 2027. The headlines scream ambition. But scroll past the press releases, and a darker signal emerges – one that directly threatens the backbone of Bitcoin's hashpower in Sichuan.
Volatility isn't a bug, it's a feature. And right now, the volatility isn't in crypto markets – it's in the power grids that mining rigs and AI data centers both depend on. The Chengdu plan is not a crypto story, but it becomes one when you trace the energy lines.
Context: Why This Matters Now
Chengdu sits in the heart of Sichuan, a province that, for years, has been Bitcoin mining's cheapest energy haven. Seasonal hydropower during the rainy season (May to October) has historically lured miners to set up shop in abandoned factories and garages. The local government previously tolerated this gray industry because it absorbed excess power. But that tolerance has a shelf life.
Enter the AI+ action plan. The policy explicitly calls for construction of massive computing infrastructure: the Tianfu Smart Computing Center, aiming for 1000 petaflops by 2027, and the Chengdu Supercomputing Center. These facilities are not optional – they are the physical chassis for the entire AI agenda. And they are power hogs.
A single H100 cluster consumes as much electricity as a mid-sized Bitcoin mining farm. The difference? AI data centers run year-round, while Sichuan's miners already compete with the dry season's higher industrial power prices. Now, the government is signaling that priority power access will flow toward AI and smart terminal manufacturing. Don't regret the dance – but the music is changing.
Core: The Unspoken Hashrate Squeeze
The policy's technical details are empty – no specific model architectures, no training framework mentions, no AI chip localization roadmap. It reads like a typical Chinese local government proclamation: big numbers, vague pathways. But that vagueness hides a brutal reality.
Based on my experience analyzing institutional energy contracts for crypto firms, here's what the policy doesn't say: The 2600 billion yuan target implies a compound annual growth rate above 30% for the AI sector. To hit that, computing power demand will double every 18-24 months. Where will that electricity come from? Sichuan's hydropower capacity is not elastic. New dam projects are frozen due to environmental regulations. The only source left is reallocating existing industrial power quotas – and Bitcoin miners hold the largest uncommitted block of baseload demand in the province.
Let me be direct: over the past three years, Sichuan's Bitcoin mining hashrate share dropped from about 15% to 9% of the global total. This policy accelerates that decline. The government has already started informal conversations with large mining farms about relocation or conversion into AI computing centers. One miner I spoke with last week told me, 'They're offering us a buyout, but the per-MWh compensation is half our current power contract.' Liquidity is vanity; solvency is sanity. For miners, solvency means energy cost control, and that control is being ripped away.
Contrarian: The Real Bottleneck Isn't AI – It's Centralization of Hashpower
Every crypto analyst fixates on Bitcoin's fourth halving and its impact on miner revenue. They point to the falling block subsidy and rising fees. But they miss the story happening on the ground in Sichuan: a government-backed power grab that will consolidate what's left of the province's mining into three or four industrial-scale players.
The policy's failure to address AI chip supply (as the Chinese analysis notes) forces firms to rely on Nvidia's black market or Huawei's Ascend. Both are energy-inefficient compared to the latest tech, meaning more watts per FLOP. More watts means more strain on the grid. The government's solution? Prioritize AI and let the 'non-strategic' industries – i.e., Bitcoin mining – fade.
But here's the contrarian twist: This isn't just about energy competition. It's about hashpower centralization. When small and mid-sized miners in Sichuan are squeezed out, the remaining hashrate will coalesce into a few large, well-capitalized pools that can afford to operate at narrow margins. Exactly the scenario I've warned about since the halving. The Chengdu policy is accelerating the very outcome that makes Bitcoin's decentralization claim hollow.
Green candles only tell half the story. The other half is written in power purchase agreements and government subsidies. And right now, Beijing is rewriting the energy narrative, district by district.
Takeaway: Watch the Floodgates
Next time you see a headline about Sichuan's 'AI miracle,' remember: the miracle runs on the same electrons that used to mint new BTC. The canary in the coal mine isn't hashrate charts – it's the price of industrial power in Chengdu. If it spikes above 0.6 yuan per kWh during the wet season, the mining exodus has begun. Don't say you weren't warned.