It began as a founding narrative of the Bitcoin maximalist era. In 2021, Jack Dorsey, then CEO of Twitter and Square, announced Block would build an open, decentralized Bitcoin mining system—a 3nm ASIC chip that would challenge Bitmain’s hegemony and democratize the supply chain. The industry cheered. Core Scientific, then one of the largest public miners in North America, placed an order. Fast forward to early 2026: Core Scientific terminated that contract, paying Block $41.9 million in cancellation fees. The chip never reached scale. The dream collapsed. And the real story isn’t about a failed product—it’s about a structural reordering of the entire Bitcoin mining economy.
What happened? The official narrative from Block was innocuous—“strategic pipeline reallocation.” Core Scientific’s CEO, Adam Sullivan, cited “a re-evaluation of capital deployment priorities.” But the numbers tell a different story. Block’s Proto chip, a 3nm beast capable of 15 exahash, was supposed to be the next-generation workhorse. Instead, Core Scientific walked away after paying an initial deposit, leaving Block with a product and no major customer. Over the past 18 months, Block’s entire crypto hardware division has been in a state of suspended animation. Its market share is negligible. The only public benchmark—a single test deployment at a Block-owned facility—showed the chip underperforming Bitmain’s S21 Pro by 12% in efficiency (J/TH). For a new entrant, that’s a death sentence.
But the deeper signal is not about Block’s engineering failure. It’s about Core Scientific’s strategic pivot—a pivot that reveals the true nature of the “Bitcoin mining industry” in 2026. Core Scientific, after emerging from bankruptcy in 2023, signed a 15-year AI data center hosting agreement with AMD. The deal is worth an estimated $140 billion in potential revenue over its lifetime. To make room, Core Scientific is converting its Texas and North Carolina mining facilities into high-performance computing centers. The $41.9 million cancellation fee is pocket change compared to the AI revenue. The message is clear: Bitcoin mining is no longer the highest-value use for land, power, and capital. AI compute is.
This is not a local event. It’s a systemic narrative shift. When a top-tier Bitcoin miner—one that survived the 2022 contagion, the 2023 miner capitulation, and the 2024 halving—decides to burn bridge with a high-profile chip supplier and divert 80% of its new power capacity to AMD, you are witnessing the end of an era. The “Bitcoin mining is digital gold mining” analogy is dead. Welcome to “Bitcoin mining is just another energy-intensive compute application—and AI wins.”
Let’s unpack how we got here. Alpha found in the noise.
Context: The Rise and Fall of Block’s Proto
Jack Dorsey’s vision for Block’s mining chip was ambitious: an open-source hardware stack, transparent supply chain, and a chip that could be mined by individuals. It was a direct counter-narrative to the Bitmain oligopoly—a narrative of decentralization through hardware. Block recruited a team of former Intel and AMD engineers, invested over $300 million, and by 2024, had a working 3nm prototype. Core Scientific, then struggling with debt and looking for a competitive edge, signed a multi-year purchase agreement for 50,000 Proto units.
The deal was announced with fanfare at the Bitcoin 2024 conference in Miami. “We’re democratizing mining,” Dorsey said. “Block is building the infrastructure for a truly decentralized network.” The market responded: Block’s stock rose 3% that day. Core Scientific’s bonds rallied.
But beneath the headlines, the technical reality was different. The Proto chip’s performance metrics were never independently audited. I’ve audited tokenomics and hardware specifications for a dozen Layer-1 projects since 2018—this smelled like a marketing campaign dressed as a product launch. The chip’s power efficiency, at 25 J/TH, was competitive with Bitmain’s S21 Pro at 23 J/TH, but the S21 was already in mass production with a 70% market share. Proto had no economies of scale, no established maintenance service, and no resale market. Core Scientific’s own preliminary tests showed that in their specific Texas environment—high ambient temperatures, variable power prices—the Proto’s real-world efficiency dropped to 30 J/TH. That’s a 33% premium over the competition. You can’t compete on hash price with a 33% power penalty.
In late 2025, Core Scientific’s board met with its restructuring advisors. The AI opportunity was knocking: a hyperscaler needed 500 MW of compute capacity. The cost of converting an existing mining site to AI was $10 million per 10 MW. The revenue multiple from AI compute was 15x higher than Bitcoin mining at current hash prices. The decision was obvious. The $41.9 million cancellation fee was a cost of doing business—a small price to shed a legacy hardware commitment and embrace the AI future.
Collapse detected. Lessons extracted.
Core Insight: The Structural Inversion of Mining Economics
Here is what most analysts miss: Core Scientific’s walkaway is not just a company’s strategic pivoting; it is a systemic signal about the long-term viability of Bitcoin mining as a stand-alone business. To understand this, we must look at the fundamental equation of mining profitability:
Mining Revenue = Block Reward + Transaction Fees – Power Cost – Hardware Capex – Cooling/OpEx
In 2024, the halving cut block rewards to 3.125 BTC. Assuming Bitcoin stays at $100,000, that’s $312,500 per block. But power costs have been rising—especially in Texas, where ERCOT wholesale prices averaged $40/MWh in 2025, up from $28 in 2023. Meanwhile, hardware costs are flat to falling. The net profit margin for a typical miner dropped from 60% in 2023 to 35% in 2025. At the same time, AI compute hosting margins are north of 70%.
Now overlay the competitive dynamics: Bitmain and MicroBT control ~90% of the ASIC market. They have the scale to drive down prices and the capital to subsidize customer acquisition. Block, as a new entrant, had no such advantage. Its Proto was a product in search of a market—and the market said no thanks. When even the biggest public miner who placed the initial order abandons you, the product is effectively dead.
But the deeper insight is about capital allocation. Core Scientific’s decision reflects a rational calculation that the return on invested capital (ROIC) for AI compute is structurally higher than Bitcoin mining for the foreseeable future. Why? Because AI compute demand is growing at 30% CAGR, while Bitcoin hash rate growth is slowing to 15% CAGR. Both compete for the same scarce resources: low-cost power, land with good connectivity, and regulatory clarity. In a world of scarce capital, the higher ROIC wins. This is not a temporary cycle—it’s a secular shift.
I saw the same pattern in 2020 DeFi summer: when liquidity farming yields exceeded traditional mining returns, capital flowed from BTC to DeFi. Today, that same capital flows from Bitcoin mining to AI data centers. The label changes; the economic principle does not.
Yield farming’s new frontier.
Contrarian Angle: The Narrative That Kaput Is Good for Bitcoin
Most Bitcoin maximalists view this event as bearish: “Mining is leaving Bitcoin; security will decline; the network is doomed.” I disagree. This is actually a constructive signal for Bitcoin’s long-term health. Here’s why.
First, the network’s security depends on total hash rate and its distribution, not on the profitability of any single miner. As miners like Core Scientific pivot, other pure-play miners (Marathon, Riot, Hut 8) will fill the gap—but at a lower cost basis. The hash rate will continue to grow, but at a slower pace, which reduces the energy intensity of the network. Over time, this leads to a more efficient mining ecosystem where only the most efficient ASICs survive. Block’s failure validates the oligopoly’s strength, but it also removes a source of hype and irrational investment.
Second, the pivot to AI provides a natural hedging mechanism for miner cash flows. Companies like Core Scientific are no longer binary bets on Bitcoin price. They have diversified revenue streams, which lowers their risk of bankruptcy and reduces the systemic risk of mass miner capitulation during bear markets. In 2022, miners defaulted on loans and dumped BTC. By 2026, miners with AI contracts can weather a 50% Bitcoin drawdown without forced selling. That’s a stabilizing force for the entire crypto market.
Third, the failure of Block’s “open mining” narrative is a healthy dose of reality. The hardware industry is a capital-intensive game with enormous barriers to entry. Attempts to decentralize it via altruistic projects almost always fail because they ignore the economics of scale. The lesson is not that decentralization is futile, but that hardware decentralization must come from modular, open-source software stacks—like Stratum V2—not from boutique chip designs. Block spent $300 million to learn what Bitmain knew for a decade: mining ASICs are a commodity, and the only differentiation is price per terahash.
Takeaway: The Next Narrative Is Compute Intermediacy
What comes next? Before you ask “which mining stock to buy,” ask yourself: “Which data center operator with mining roots has the best AI contract pipeline?” Core Scientific’s AMD deal is just the beginning. Expect Marathon Digital, Riot Platforms, and Hut 8 to announce similar partnerships within the next six months. The lines between “Bitcoin miner” and “AI cloud provider” will blur. The market will price these companies as infrastructure plays, not crypto plays.
Block’s mining chip saga is a footnote in that bigger story. Jack Dorsey’s vision was noble but premature. The real alpha is in the convergence of crypto-native capital with AI-native compute demand. The noise of a contract cancellation obscured the signal: Bitcoin mining is growing up. It’s leaving the rebellious teenager phase and entering the boring, profitable middle age of a utility provider.
So, when you read headlines about Block’s failed chip, don’t mourn. Question the narrative. The collapse of an aspirational product teaches us exactly how the mining industry will evolve: efficiently, ruthlessly, and aligned with the highest bidder for electrons. That’s not a loss—it’s maturation.