The shareholder vote was a binary choice: take $9 billion in cash, or gamble on a partnership with AMD. They chose the gamble. The ledger keeps score.
Context: Core Scientific, a Nasdaq-listed Bitcoin miner (CORZ), is pivoting to AI data center hosting. The AMD partnership, announced as a strategic move, is supposed to justify rejecting the $9B acquisition offer. But the announcement is a press release, not a technical document. The market swarmed with optimism. I see a different picture.
Core: The Mechanical Reality Check
Let’s dissect the AMD partnership. The official statement mentions “collaboration” and “deployment of AMD Instinct GPUs.” That’s it. No power capacity numbers. No committed MW. No test results. No ROCm vs. CUDA benchmarks. The article hypes the deal as transformative, but the data is absent.
From my years auditing mining operations, I know that converting a Bitcoin mine to an AI datacenter is not plug-and-play. You need liquid cooling, high-density racks, InfiniBand networking, and GPU cluster orchestration. AMD’s ROCm software stack, while improving, still lags behind Nvidia’s CUDA in maturity and ecosystem support. The article doesn’t address these engineering bottlenecks. It’s a strategic announcement, not a technical one. Minted nothing, promised everything.
The capital structure adds another layer of risk. Core Scientific emerged from bankruptcy in 2023 with debt. The AI pivot requires massive CapEx. The article implies the partnership alone will fund the transition, but there’s no mention of equity dilution or debt financing. The shareholders rejected $9B in cash, effectively setting a valuation floor. Now the market will watch every earnings report to see if the company can generate more than that. The power bill doesn’t lie. The ledger keeps score.
Contrarian: What the Bulls Got Right
I’m not here to dismiss the entire thesis. The bulls have a point: Core Scientific’s existing power infrastructure is a real asset. Long-term power purchase agreements (PPAs) locked in at low rates give them a cost advantage over pure-play AI cloud providers. AMD needs deployment partners to challenge Nvidia’s dominance. This isn’t just a supply deal; it’s potentially a joint engineering collaboration. AMD wants real datacenter data to optimize their Instinct GPUs for AI workloads. Core Scientific provides that proving ground.
The article correctly notes that the rejection of the $9B sale signals a conviction that the company is worth more. Under the right execution, the AI hosting business could generate higher multiples than mining alone. The infrastructure reuse model is economically sound—if they can solve the technical challenges.
Takeaway: The Real Test
But conviction is not execution. The next 12 months will reveal whether the AMD partnership is a revenue driver or a marketing prop. The first concrete milestone will be the disclosure of delivered MW capacity and utilization rates. If the next quarterly report shows no meaningful revenue from AI hosting, the market will reprice the stock. The shareholders bet $9B on a promise. The ledger keeps score, and it will settle that bet soon enough.
Until then, this is a story of intent, not code. And in this industry, intent is fiction.
