Hook
Over the past week, a single transaction has been haunting my on-chain screens. Not a whale moving 10,000 BTC, nor a DeFi exploit draining a protocol. It’s a corporate wire transfer: $41.9 million paid by Core Scientific to Block, just to break a contract for 15 exahash of bitcoin mining chips. That’s roughly 3% of the entire Bitcoin network’s current hash rate — turned into a penalty. When a company willingly burns cash to avoid receiving hardware, the data screams one thing: something is deeply wrong.
I’ve seen this pattern before. In 2017, I tracked wallets for ICOs that looked promising until insider addresses started dumping pre-launch. The on-chain evidence was clear: the project had lost its backbone. Today, the backbone isn’t a wallet — it’s an order book. And the $41.9M exit fee is the on-chain rumor that demands investigation.
Context
To understand the gravity, let’s rewind. Block, Inc. — the company led by Jack Dorsey — launched its Proto mining chip initiative in 2023 with fanfare. The pitch: a 3nm ASIC that would challenge Bitmain’s dominance. Core Scientific, then one of America’s largest publicly traded miners, signed on as the flagship customer, ordering 15 EH/s worth of units. The deal was supposed to start deliveries in early 2024.
But by the first quarter of 2025, Core Scientific filed a termination notice, paying $41.9 million to walk away. The reason? A “strategic pivot” away from bitcoin mining toward AI infrastructure. In fact, Core simultaneously announced a massive contract with AMD to lease 200 MW of data center capacity for high-performance computing, projecting $14 billion in revenue over 15 years.
That pivot is the context we need to dissect. It’s not just about one chip order. It’s about capital, energy, and talent flowing from the Bitcoin mining ecosystem into the AI gold rush. And it raises an uncomfortable question: Are we witnessing the commoditization of Bitcoin mining, or just a bad product launch?
Core: The On-Chain Evidence Chain
Let’s treat this like a data detective would — trace the clues.
Clue #1: The 15 EH/s That Never Mined. Core Scientific’s termination means Block now has a factory of 3nm chips with no home. In the mining hardware market, 15 EH/s of capacity is a serious supply shock. If Block dumps these chips on the secondary market at a discount, it could depress used ASIC prices across the board. I’ve seen this happen in DeFi: when a large LP exits a pool ungracefully, the impermanent loss cascades. The on-chain analogue here is a sudden flood of supply — but for physical hardware.
Clue #2: The $41.9M as a Signal of Mispricing. Core Scientific paid that fee rather than take delivery. That implies the chips’ expected revenue over their lifetime was less than the penalty. Let’s do rough math: a 15 EH/s order, assuming 25 J/TH efficiency and $0.05/kWh power, would generate roughly $120 million in gross revenue over four years at current Bitcoin prices. But after subtracting electricity, maintenance, and the cost of the hardware itself, the net profit might have been razor-thin. If Core’s internal models showed a negative NPV even after paying the cancellation fee, Block’s chip was not competitive.
I recall my DeFi Summer tracking days: when a new liquidity pool boasted 50% APY but attracted only small whales, I knew the yield was unsustainable. Here, the yield from Block’s chips simply didn’t match the risk.
Clue #3: The Broader Block Crypto Graveyard. This mining chip failure is not isolated. Let’s line up the data: - Tidal (music streaming) — acquired for $300M, later sold at a loss. - TBD (decentralized identity/Web5) — shut down after minimal traction. - Bitkey (self-custody wallet) — low adoption, effectively dead. - Bitchat (P2P messaging) — never launched beyond beta. - Cash App (payment) — fined over $200M by CFPB and state regulators for fraud handling.
In the past five years, Block’s stock has fallen 68%. From ICO chaos to crystalline clarity, this is a classic case of a company overextending its reach into hardware without the supply-chain moats of Bitmain or MicroBT.
Clue #4: The Capital Rotation to AI. Core Scientific’s move is a smoking gun. In the same quarter they terminated the Block contract, they announced a $14B AI compute deal with AMD. That’s not a pivot — it’s a migration. Mining companies own land, power, and cooling infrastructure that AI data centers need. The incentive to shift is enormous: AI compute margins are 3-5x higher than mining margins at current rates.
I’ve been tracking wallet flows on Nansen for years. When a whale starts moving ETH from DeFi protocols to centralized exchanges, you know they are preparing to exit. Here, Core Scientific is moving its capital from mining hardware orders to AI infrastructure leases. The signal is loud: merchant mining is becoming a lower-priority use case for industrial power.
Clue #5: The Hash Rate Plateau Risk. Bitcoin network hash rate has grown steadily, but if large miners like Core shift capacity to AI, the growth rate may slow. This doesn’t mean Bitcoin breaks — it just means the security budget becomes more dependent on a smaller number of dedicated miners. And if those miners struggle to compete with AI for power, transaction fees may need to cover a larger share of security costs.
Contrarian: Correlation ≠ Causation
Now, let me challenge my own narrative. Is Block’s chip actually bad, or is Core’s pivot a strategic bet that could backfire?
The Chip Might Be Fine. We haven’t seen third-party benchmarks. The 3nm process is cutting-edge — TSMC’s N3 offers 30% better efficiency than Bitmain’s 5nm chips in some designs. Block could have produced a competitive product. Core’s exit might not be about chip performance but about opportunity cost: they saw a massive AI deal and had to free up capital. The $41.9M penalty might be cheaper than missing the AI window.
The AI Revenue Is Speculative. Core’s $14B projection assumes AMD’s GPU demand remains high for 15 years. That’s a long horizon in tech. If the AI hype cycle cools, Core could be left with empty floors and massive debt. Meanwhile, Bitcoin mining, though volatile, has a proven long-term revenue model through block rewards and fees.
Whales Don’t Hide; They Just Swim in Deeper Waters. The biggest Bitcoin miners — Bitmain, MicroBT, Riot Platforms — are still expanding hash rate. Riot recently announced a new 1 GW facility in Texas. The fact that one miner pivoted doesn’t mean the entire industry is dying. It means the low-hanging fruit of cheap power is being claimed by AI, but mining will continue at scale where electricity is cheapest (stranded gas, hydro, etc.).
Block’s Failure Could Be a Buying Opportunity. Block’s stock is beaten down. If the market has already priced in the crypto failures, the real value might be in its core payment business (Cash App/Square). The mining chip debacle might be a one-time write-off, not a systemic rot.
So the contrarian view is: Don’t extrapolate one termination into a death knell for Bitcoin mining. Core’s move is a valid capital allocation decision, not a verdict on Block’s product. The truth likely sits in the middle — the chip was decent but not game-changing, and the AI attraction was too strong to ignore.
Takeaway: The Signal to Watch Next Quarter
Parsing the noise to find the signal’s heartbeat, I’ll focus on three data points over the next 90 days:
- Block’s Q2 2025 earnings call — Listen for any mention of Proto’s future. If they announce a pivot or write-down, the signal is confirmed. If they claim to have found new buyers, the story changes.
- Core Scientific’s AI revenue line — If they report even a few million dollars from the AMD contract in Q2, the pivot gains credibility, and other miners may follow.
- Secondary ASIC market prices — If Block’s dumped chips flood the market, expect used S19 and M50 prices to drop 10-15%. That would be a buying opportunity for small miners but a squeeze for hardware vendors.
From ICO chaos to crystalline clarity, one thing is certain: the era of “build a chip and they will come” is over. Bitcoin mining has matured into a hyper-competitive, capital-intensive industry where even a Jack Dorsey can stumble. Eyes wide open, data streams wide — the next step isn’t about who has the best vision, but who executes the best supply chain.
***