A company that just clawed its way out of bankruptcy rejects a $9 billion acquisition. Days later, it announces a partnership with AMD. The stock jumps. The crypto Twitter crowd cheers. But the data detective sees a different story: the numbers don't add up yet.
Core Scientific trades on Nasdaq as CORZ. It runs Bitcoin mining rigs and, increasingly, AI data centers. On paper, the pivot from burning electricity for SHA-256 to powering GPUs for large language models is elegant. Both businesses consume megawatts. Both require low-cost power and high-density infrastructure. The synergy is obvious. But the execution gap is vast.
Let me be clear: I am not here to dismiss the thesis. I am here to dissect it. Based on my experience auditing 15 ICOs in 2017, I learned that strategic announcements without technical milestones are noise. The alpha isn't in the press release; it's in the silenced code. And in this case, the code is missing.
Context: The Two-Legged Stool
Core Scientific's business rests on two legs. The first is Bitcoin mining. Post-halving, that leg is bleeding. The fourth halving slashed block rewards by 50%, and hash price has dropped to all-time lows. Miners with inefficient rigs or high power costs are being squeezed out. Hash power is concentrating into the top three pools. The decentralization consensus is hollowing out. Core Scientific, with its legacy fleet of ASICs, is not immune.
The second leg is AI hosting. In 2024, the company signed multi-year contracts with CoreWeave to host GPUs for AI workloads. This is the growth story. The AMD partnership, announced in early 2025, is supposed to accelerate that. AMD will supply Instinct GPUs to Core Scientific's data centers. The market interpreted this as a direct challenge to Nvidia's dominance in AI hardware.
But here is the problem: the announcement contained zero technical details. No number of GPUs. No megawatts of capacity. No timeline for deployment. No software stack benchmarks. No contract value. Nothing. The press release is a marketing document, not a technical specification.
Core: The Evidence Chain
Let me build the on-chain evidence chain, even though this is an equity, not a crypto token. The relevant data is not on-chain; it is in the company's financial filings and operational metrics. But the same detective mindset applies.
First, the capital structure. Core Scientific emerged from Chapter 11 bankruptcy in January 2024. The restructuring left legacy debt. The exact terms are not public, but the company's balance sheet is fragile. The AI pivot requires massive capital expenditure: liquid cooling retrofits, high-density power distribution, InfiniBand or RoCE networking, and GPU purchases. The AMD partnership could involve favorable financing or joint engineering, but the press release didn't confirm that. The risk of equity dilution is real. If Core Scientific issues new shares to fund the pivot, existing shareholders' stake gets diluted. The $9 billion offer rejected by shareholders might have been a better exit than a diluted future.
Second, the technical challenge. Converting a Bitcoin mining facility to an AI data center is not plug-and-play. Bitcoin ASICs are air-cooled, low-maintenance, and tolerate wide temperature ranges. GPUs for AI training require liquid cooling, precise climate control, and high-speed interconnects. The networking layer alone is a different universe: mining uses simple stratum protocols; AI clusters use InfiniBand or RoCE v2 with nanosecond latency requirements. Core Scientific's operational expertise is in mining, not in high-performance computing. The learning curve is steep.
Third, the software ecosystem. Nvidia's CUDA platform is the default for AI development. AMD's ROCm is playing catch-up. Many AI frameworks (PyTorch, TensorFlow, JAX) have CUDA-optimized kernels. ROCm support is improving, but it lags in performance and stability for large-scale training. Core Scientific will need to ensure its AMD GPUs deliver competitive performance. If they don't, customers will go elsewhere. The partnership does not guarantee adoption.
Based on my experience developing a DeFi arbitrage script in 2020, I know that the difference between a profitable trade and a loss is often hidden in the latency of data feeds. Similarly, the difference between Core Scientific's success and failure is hidden in the execution details of the AMD partnership. And those details are not public.
Contrarian: The Correlation Trap
The market is correlating the AMD partnership with a higher valuation. Shareholders rejected $9 billion, implying they believe the company is worth more. But correlation is not causation. The assumption that the AI pivot will succeed is priced in, but the risks are not.
Consider the counterfactual: if the AI hosting market cools, or if AMD's GPUs underperform, or if Core Scientific's power contracts are not competitive, the company could be worth significantly less than $9 billion. The Bitcoin mining leg is already declining. Hash rate concentration is a systemic risk. If the AI pivot fails, the company has no fallback. The rejection of the $9 billion sale could become a catastrophic error.
Scarcity is an algorithm, not a belief system. The value of Core Scientific's shares will be determined by the algorithm of revenue, margins, and growth. Belief in the AMD partnership is not enough. The data must speak.
I also see a hidden signal: AMD's motivation. AMD needs real-world deployment data to validate its Instinct GPUs against Nvidia's. Core Scientific provides that. The partnership might be a joint engineering effort, not just a GPU purchase. If so, the value is in the co-optimization, not in the hardware. But that is speculative. The press release did not confirm it.
Takeaway: The Next 90 Days
Here is what I will be watching. The Q3 2025 earnings release should include: - Number of megawatts of AI capacity delivered. - Utilization rates of those data centers. - Revenue per megawatt from AI hosting vs. mining. - Any update on the AMD GPU deployment timeline.
If those metrics are positive, the thesis gains credibility. If they are absent or weak, the rejection of $9 billion will look like a mistake. The ledger remembers what the marketing forgets.
Correlations are the lie; liquidity is the truth. The liquidity of Core Scientific's stock is thin. Institutional investors are waiting for proof. The data detective will be watching the quarterly filings, not the hype.
Due diligence is the only hedge against chaos. And right now, the due diligence on Core Scientific's AMD partnership reveals more questions than answers. The market is betting on a future that has not been built. I prefer to wait for the data.