Over the past quarter, a single transaction whispered louder than any price chart. Core Scientific paid $41.9 million to cancel orders for Block’s Proto mining chips. Not a negotiation. Not a delay. A full termination. The number—equivalent to the net profit of a mid-tier mining pool over six months—was buried in an SEC filing, but its implications ripple through every hash board and power purchase agreement in the industry. I watched this unfold from my desk in Doha, and the setup told me something the headlines missed: this wasn’t just a failed product launch. It was a structural pivot away from Bitcoin mining as we know it.
Context is everything. Block, led by Jack Dorsey, poured years and millions into developing a 3-nanometer mining chip capable of delivering 15 exahashes per second. The Proto chip was supposed to challenge Bitmain and MicroBT, two giants that control over 80% of the ASIC market. Core Scientific was the only public major customer. They signed on in 2024, promising to deploy thousands of units. By 2025, they were paying to walk away. Block’s chip was not a technical failure—it was a commercial one. The company couldn’t match the cost-efficiency of established competitors, and Core realized that deploying those machines would yield negative returns at current hash rates and Bitcoin prices.
But the story runs deeper. Core Scientific didn’t just cancel the order; they redirected their capital into a $14 billion, 15-year contract with AMD to host AI computing infrastructure. They are turning their mining data centers into AI data centers. This is the core of the analysis: the resources that once powered Bitcoin’s security—land, power, cooling, capital—are now flowing toward artificial intelligence. The math is brutal. A single AI rack can generate 10x the revenue of a mining rack per megawatt. The market is pricing this shift fast. Core Scientific’s stock has rallied as they announce the pivot, while Block’s stock is down 68% over five years, weighed down by a string of failed crypto ventures: Tidal, TBD, Bitkey, and now the chip.
I see the order flow every day. Mining equipment orders are declining. Secondary market prices for ASICs are sliding. Hashrate growth is slowing from 50% annual rates to below 20%. Meanwhile, power contracts that once went to miners are being snapped up by AI hyperscalers. Holding the line when the world screams to sell—that was my rule during the 2022 bear market, and it applies here. But the line has shifted. The line is no longer about Bitcoin price; it’s about whether mining can compete with AI for the same finite resources.
The contrarian angle is this: the market is celebrating Core Scientific’s pivot as a brilliant strategic move, but it’s actually a desperate survival tactic. The $41.9 million termination fee is a sunk cost, a small price to escape a losing bet. The real risk is that AI infrastructure buildout becomes overcrowded. Every major mining company is now exploring AI hosting. Marathon, Riot, Hut 8—all are talking about AI. If supply of AI-ready data center space surges, rental yields compress. Core’s $14 billion contract with AMD is not guaranteed revenue; it’s a maximum potential over 15 years, with cancellation clauses. The same risk of overcapacity that plagued mining is now entering the AI hosting space. The crowd is rushing to the new narrative, and that often marks the top of the opportunity.
For Block, the failure of Proto is not the end. The chip design and IP have salvage value. A distressed buyer could acquire the technology at a discount and deploy it in niche markets—off-grid mining in stranded energy sites, for example. But that’s a low-probability play. The more likely outcome is that Block writes down the entire project, closing a chapter on Dorsey’s crypto hardware ambitions. The signal for traders is clear: avoid mining-exposed equities that lack an AI pivot, and watch for the moment when the AI frenzy itself cools.
The takeaway is actionable. Holding the line when the world screams to sell means not buying the AI pivot story at its peak. Instead, wait for the shakeout. Identify mining operations with strong balance sheets and diversified revenue—those that can survive two years of compressed margins. The hash ribbon indicator shows that miner capitulation is increasing. When the last of the marginal miners exit, the survivors will capture a higher share of the block rewards. That is the entry point. Holding the line when the world screams to sell is not passive—it’s a disciplined wait for the structural bottom.
Feel the trend, but don’t chase it. The trend is resource migration from Bitcoin to AI. The chase is buying Core Scientific at 10x earnings. The profit is in the pause—the quiet accumulation of assets that others are fleeing. Check your positioning. The $41.9 million signal is not a warning; it’s a confirmation that the old mining playbook is obsolete. Rewrite yours.