The Pickaxe Premium Is Over: Why Yang Zuoxing Just Declared the End of Bitcoin Mining‘s Golden Era

I didn‘t need Yang Zuoxing to tell me the party was over. I saw it in the order books. The smell of burnt-out ASICs and the silent auction for power allocation. The man who co-founded MicroBT, the architect of the Whatsminer empire, stood on a stage in July 2026 and said what every battle-hardened trader in this space already felt: the golden age of Bitcoin mining is dead. Not dying. Dead. The blockchain doesn't lie. The numbers are right there in the P&L of every manufacturer. The pickaxe premium—the insane margin that comes from selling shovels during a gold rush—has evaporated. For three mining cycles, from 2017 through 2025, the total sales value of mining rigs stubbornly parked itself between 300 and 400 billion yuan. A flat line. No growth. Meanwhile, the gross margin on those rigs collapsed from a fat 80-90% down to a gut-punching 20-30%. That‘s not a cyclical dip. That is a structural rupture. Yang called it the "end of the golden era." I call it the moment the market flips from a growth story to a survival story. And the survivors won’t be the ones with the most hashrate. They‘ll be the ones who understand the new energy physics. The context is brutal but simple. Bitcoin mining is a commodity business masquerading as a tech business. For years, the narrative was about chip efficiency—nanometers, J/TH, next-gen ASICs. The reality? Efficiency gains have hit diminishing returns. The low-hanging fruit is gone. The real battlefield has shifted from the foundry to the substation. It’s not about who has the best 3nm design anymore. It‘s about who can secure the cheapest electrons. The halving in 2024 slashed block rewards by 50%, and AI is now hoovering up both capital and power. Yang’s three escape hatches—natural gas flare mining, AI-integrated data centers, and solar mining—aren‘t just R&D projects. They are survival parachutes for a industry that realizes its core product (a centralized hash rate) is facing a margin squeeze that, if unaddressed, will lead to massive miner capitulation. I don’t think the average crypto Twitter bro understands this. They see BTC at $100k and think mining is printing money. They don’t see that the cost to print that money has doubled while the sale price of the printing press has stagnated. Let‘s look at the core of this: the data. The table is sickening. 2017-2020 cycle: total sales held flat at ~300 billion yuan, margins ~85%. 2021-2024: sales nudged up to ~350 billion, margins dropped to ~50%. 2025 onward: sales stuck at ~400 billion, margins cratered to ~25%. That’s a 60 percentage point margin compression. What does that mean in real terms? The mining rig business used to be the most profitable hardware business on earth. Now it struggles to match the margins of a mid-tier laptop manufacturer. The pickaxe premium is gone. Why? Because the number of ASICs shipped hasn‘t grown proportionally to the total sales value. Inflation and higher BTC prices masked the unit decline. But the margins scream the truth: there are too many players, too much capacity, and not enough differentiation. The moat was chip design, but that moat has been breached by competition. The real moat now is access to stranded energy—natural gas that would otherwise be flared, solar power at remote locations, or waste heat reuse. That’s why Yang‘s talk of AI integration is so critical. It’s not about making mining rigs smarter. It’s about repurposing the infrastructure asset—the power, the cooling, the real estate—to serve AI inference workloads. That shifts the revenue model from single-purpose (hashrate) to multi-purpose (hashrate + AI compute). If that works, the ASIC manufacturer becomes an infrastructure provider. If it fails, the industry shrinks to a long tail of hobbyist miners. Now, the contrarian angle. The market narrative is hopeful. "Natural gas mining will save us!" "AI integration will double the value of our data centers!" "Solar mining is the future!" I call hopium. The blockchain doesn‘t forgive bad capital allocation. Let’s dissect each: Natural gas flare mining is a well-known, niche play. It works in the Permian Basin and a few other oil fields. But scaling it globally requires cheap gas in places where you can also get cheap fiber and stable governance. That‘s a tiny intersection. Solar mining faces the intermittency problem—you need batteries, and batteries add cost. The economic viability of solar mining is only positive if the BTC price is high enough and the mining equipment is already depreciated. It’s not a greenfield solution. AI integration? That‘s the most exciting, but also the most technically risky. You’re asking a hardened ASIC miner to also act as an AI accelerator. That requires custom silicon, new firmware, and partnerships with inference providers. The complexity is high. The chance of failure is non-trivial. The real contrarian take is this: the "golden era" ending is actually healthy. It forces capital discipline. It weeds out the weak. The miners that survive will have the lowest cost of energy and the most efficient operations. The industry will consolidate. The ASIC manufacturers that pivot to AI infrastructure will create new revenue streams. But most won’t. Most will bleed out slowly. I‘ve seen this play out in every commodity cycle. The moment margins compress, the smart money exits the hardware business and buys the underlying asset. In this case, that might mean buying BTC directly rather than mining it. Or buying the power generation assets that serve mining. The miners themselves? They’re sitting on depreciating assets with negative carry. Not a fun place to be. Takeaway: actionable levels. The hashprice—the average revenue per TH/s per day—is the single metric to watch. If it falls below $50/PH/s/day for an extended period, expect a cascade of miner bankruptcies and a hashrate drawdown of 20-30%. That‘s the blood in the water. On the ASIC side, any new rig priced above $20/TH is a value trap unless accompanied by an AI integration roadmap. The price of second-hand S19s will tell the story: if they trade below $5/TH, the industry is in full retreat. I don’t have a magic crystal ball. But I have a PhD in cryptography and a gut fed by six years of watching this dance. Golden ages don‘t return. They’re replaced by gritty ages. The question is whether the grit is enough to build something new.