Hype is noise. Standards are signal.
Shareholders just rejected a $9 billion buyout. Days later, the company announced a partnership with AMD. The stock jumped 12% on the news. The narrative is clear: Core Scientific is transforming from a Bitcoin miner into an AI data center powerhouse. But I’ve audited 15 DeFi protocols during the 2020 summer. I’ve seen partnerships that were nothing more than press releases. This one smells the same.
Let me show you what the headlines missed.
Context: The Core Scientific Story
Core Scientific is one of the largest Bitcoin mining operations in North America. They emerged from Chapter 11 bankruptcy in 2023, carrying legacy debt but also a massive infrastructure advantage: long-term power purchase agreements at below-market rates. In 2024, they signed multi-year hosting contracts with CoreWeave, a pure-play AI cloud provider. That pivot was real. It was backed by engineering work.
Then came the $9 billion acquisition offer. An unnamed buyer—likely a consortium of institutional investors—wanted to take the company private. Shareholders voted no. The official reason: the offer undervalued the company’s AI potential. The unofficial reason: the board wanted to bet on AMD.
On paper, the AMD partnership is a breakthrough. Core Scientific will deploy AMD Instinct GPUs in its data centers, offering AI compute as a service. The market cheered. But let’s dig into the fine print.
Core: What the AMD Partnership Actually Says
First, the disclosure. The partnership announcement contained zero technical details. No number of GPUs. No megawatts of deployed capacity. No performance benchmarks. No revenue commitments. No minimum purchase obligations. Nothing.
In my 29 years of observing tech markets, I’ve learned one rule: if a deal is real, the numbers come out. If a deal is a story, the PR team writes the script. This is a story.
Technical Reality Check
Converting a Bitcoin mining facility to an AI data center is not plug-and-play. Mining rigs are ASICs. They run on air cooling, low-density racks, and simple networking. AI clusters require:
- Liquid cooling for high-density GPU racks (up to 40kW per rack vs 2kW for miners)
- InfiniBand or RoCE networking for GPU-to-GPU communication
- Specialized software stacks: CUDA (Nvidia) or ROCm (AMD)
- 24/7 uptime SLAs with 99.99% availability
AMD’s ROCm ecosystem is growing. But it is not CUDA. Every operator I’ve spoken to says the same thing: the software maturity gap is at least 12–18 months. Core Scientific will need to build internal expertise from scratch. They are a mining company, not a hyperscaler.
Data from the Field
Over the past 12 months, I’ve tracked the conversion progress of five major Bitcoin miners. Only one—Hut 8—has successfully deployed high-density GPU clusters. The others are still in pilot phases. Core Scientific’s existing CoreWeave contract is for Nvidia GPUs, not AMD. The AMD partnership is a parallel track, not a replacement.
The Real Asset: Power Contracts
The real value in this deal is not the chips. It’s the power. Core Scientific holds long-term, fixed-price power agreements that are increasingly rare in the US. As AI demand soars, energy costs rise. A miner with locked-in cheap power has a structural advantage. But that advantage is only realized if the GPUs are actually running and generating revenue.
I’ve seen this movie before. In 2021, I led the Proof of Origin initiative that authenticated 5,000 NFTs. We had partnerships with marketplaces, artists, and collectors. But without a working API and cross-chain verification, those partnerships were just logos on a website. Core Scientific’s AMD deal is a logo. The real work hasn’t started.
Risk Quantification
Let me put this in numbers. Assume Core Scientific wants to deploy 100 MW of GPU capacity. At current market rates, that’s approximately $300–$400 million in capital expenditure for GPUs, networking, and cooling retrofits. The company’s market cap is around $3 billion. They have debt from bankruptcy. They will likely need to raise capital through equity or convertible notes. Dilution is coming.
The AMD partnership does not provide a committed revenue stream. It’s a supply agreement. AMD sells chips. Core Scientific buys them. If demand for AI compute softens, or if Nvidia releases a more cost-effective product, Core Scientific is left holding inventory. The risk is asymmetric: potential upside is capped by competition, but downside is real capital loss.
Contrarian Angle: The $9 Billion Rejection Was a Mistake
Here’s the take that no one wants to hear: shareholders might have made the wrong call.
A $9 billion offer for a company that just emerged from bankruptcy and is pivoting into an unproven market is a generous premium. At that price, the buyer was paying for optionality. Core Scientific’s board rejected it, betting they could create more value through the AMD partnership. But the partnership is not a guarantee. It’s a bet on execution, market timing, and AMD’s ability to compete with Nvidia.
Verify everything. Trust the protocol.
In 2022, during the Luna crash, I personally deployed $5 million to stabilize three lending protocols on Avalanche. I did not rely on press releases. I audited the smart contracts, checked the collateral ratios, and executed a rebalancing algorithm. That’s what real risk management looks like.
Core Scientific’s investors are flying blind. They have no technical validation of the AMD partnership. They have no public data on utilization rates or revenue per GPU. They are betting on a narrative, not a protocol.
Compliance is the new crypto currency.
If this were a DeFi protocol, we would demand a security audit. We would check the code. We would verify the team’s claims. But because it’s a public company, investors assume the due diligence is done by the board. That’s a dangerous assumption. The board just rejected a $9 billion offer. They are incentivized to justify that decision. The AMD partnership is their justification.
Structure wins. Chaos loses.
In my 2025 Vancouver Framework, I worked with provincial regulators to standardize compliance for $50 billion in institutional crypto assets. The key lesson: structure and transparency are prerequisites for trust. Core Scientific’s announcement lacks both. No structured data. No transparent milestones. Just a press release and a rising stock price.
Takeaway: What to Watch For
Over the next six months, ignore the stock price. Watch three things:
- Actual MW deployed: Core Scientific should report how many megawatts of GPU capacity they have energized. If they don’t, assume the partnership is still in the planning stage.
- Revenue per GPU: The company should disclose AI hosting revenue separately from mining revenue. Any obfuscation is a red flag.
- AMD’s Instinct sales: AMD’s earnings calls will reveal if they are shipping enough GPUs to meet demand. If AMD is supply-constrained, Core Scientific’s deployment will be delayed.
If these metrics are positive, the stock deserves a premium. If not, the $9 billion offer will be remembered as the one that got away.
Final thought
This is a classic inflection point. The market is betting on transformation. But transformation requires execution, not announcements. As an evangelist for decentralization, I believe in the potential of reusing mining infrastructure for AI. But I also believe in data, audits, and transparent governance.
Core Scientific’s shareholders have made their choice. Now they must demand accountability. Otherwise, the only thing being mined is hype.
