When 70% of Core Scientific's shareholders voted down the $9 billion acquisition offer, the market did a double take. Here is a public company that emerged from bankruptcy in 2023, saddled with legacy debt, and now it was turning down a cash exit that would have paid a premium over its trading price. The conventional wisdom said: take the money, restructure, move on. But the shareholders saw something else. They saw a mining company that had quietly transformed into an AI infrastructure play, and they believed the AMD partnership—announced just weeks before the vote—was worth more than $9 billion. I have spent the last 16 years watching market narratives shift from hype to reality, and this moment feels eerily familiar. It is the same pattern I saw in 2017 when Ethereum projects were valued on whitepapers alone, and in 2020 when DeFi yields masked oracle vulnerabilities. Core Scientific is not a smart contract protocol; it is a physical infrastructure company. But the market is treating it like a narrative stock. And that is where the danger—and the opportunity—lies. Trust is the only asset that survives the crash, and in this case, the trust is being placed not in audited code, but in a CEO's ability to convert megawatts of mining power into AI compute clusters. Let me be clear: I am not here to tell you whether the stock will go up or down. I am here to dissect the technical assumptions behind the bet, because every scar in the market teaches a new rule. The rule here is simple: strategic partnerships are not infrastructure deliveries. And until I see deployed GPU capacity, power utilization rates, and software stack benchmarks, the AMD deal remains a promise, not a proof. This article is a deep dive into the engineering, the economics, and the hidden risks of Core Scientific's pivot. I will walk you through the same forensic framework I use when auditing a DeFi protocol—except here, the 'code' is the physical plant, and the 'oracle' is the power grid.
Context: The Mining-to-AI Pivot Core Scientific started as a Bitcoin miner, one of the largest in North America. It owned vast facilities with cheap power contracts, often locked in for years at rates below market. When Bitcoin mining margins compressed after the 2024 halving, the company faced a choice: continue as a pure miner with declining revenue, or repurpose its infrastructure for the AI boom. AI data centers need exactly what mining farms have: massive power, cooling systems, and high-density racks. But the similarity ends there. Mining rigs are ASICs—simple, energy efficient, and designed for a single task. AI workloads require GPUs, specifically Nvidia's H100 and its successors, which demand liquid cooling, InfiniBand networking, and clusters of thousands of interconnected chips. Converting a mining facility is not a plug-and-play operation. It requires ripping out ASIC shelves, installing liquid cooling loops, upgrading the electrical distribution to handle variable loads, and building a network that can handle the east-west traffic of distributed training. Core Scientific already had a partnership with CoreWeave, a cloud AI provider, and had signed multi-year hosting contracts. But the AMD partnership, announced in early 2025, signaled a new direction: instead of just hosting Nvidia GPUs, Core Scientific would now deploy AMD's Instinct MI300X accelerators. This is a strategic bet on AMD's ability to challenge Nvidia's dominance. The shareholder vote to reject the $9 billion acquisition effectively said: we believe the AMD partnership will generate more long-term value than a cash exit. That is a bold claim, and it requires a close look at the technical feasibility and the market dynamics.
Core: The Order Flow Analysis Let me start with the numbers that matter. The original article disclosed no technical metrics—no megawatts of AI capacity, no GPU count, no expected delivery dates. That is a red flag. In my experience auditing DeFi projects, when a protocol announces a 'strategic partnership' without a concrete timeline or verification mechanism, it is often a placeholder for future funding. The same applies here. Core Scientific's shareholders are betting on a future revenue stream that is not yet contractually secured. The order flow analysis I want to focus on is the supply chain and the software stack. AMD's Instinct GPUs are competitive on paper, but they suffer from a software ecosystem gap. Nvidia's CUDA is the industry standard for AI training, and while AMD's ROCm has improved, it still lacks the libraries, documentation, and community support that CUDA offers. For a company like Core Scientific that is targeting multi-tenant AI hosting, compatibility matters. If a customer wants to train a large language model, they will likely use PyTorch or TensorFlow, which are optimized for CUDA. ROCm compatibility is often an afterthought. This means Core Scientific may face a chicken-and-egg problem: they need customers to adopt AMD hardware, but customers are reluctant to switch without proven performance. The contrarian angle here is that the market is pricing the AMD partnership as a certainty, but the engineering reality is that it takes months to validate a new GPU cluster. Based on my experience in 2020, when I helped my community exit a Curve pool before oracle manipulation hit, I learned that the most dangerous time is when everyone is optimistic but no one has tested the system under load. Core Scientific has not published any benchmark results for AMD GPUs in their facilities. They have not disclosed the power consumption per card or the cooling efficiency. Until they do, the partnership is a narrative, not a technical milestone.
Contrarian: The Retail vs. Smart Money Divide Here is where the story gets interesting. The shareholder vote was a retail-driven rebellion against a board that wanted to sell. The 'smart money'—institutional investors and activist funds—likely supported the acquisition because it offered a guaranteed premium. But retail shareholders, many of whom are crypto enthusiasts who bought the stock as a proxy for AI, believed the AMD deal would create more value. This is a classic retail vs. smart money divide, and I have seen it play out in DeFi many times. In 2022, when Terra Luna was collapsing, the community initially defended the protocol because they believed in the narrative. They ignored the on-chain data showing the algorithmic stablecoin was losing its peg. Retail investors are driven by hope and social proof; smart money is driven by risk-adjusted returns. In this case, the smart money may be right to be cautious. The AMD partnership is not a guaranteed revenue stream; it is a supply agreement. Core Scientific has to buy the GPUs, install them, and then find customers to use them. Without a guaranteed off-take contract, the company bears the capital expenditure risk. Moreover, the cost of converting a mining facility is high. A typical 100 MW data center can cost $100 million to $200 million to retrofit for AI. If the AMD partnership does not generate the expected demand, Core Scientific could be left with stranded assets—GPUs that are too power-hungry for mining and too niche for general cloud. The shareholder rejection of the $9 billion acquisition essentially said: we trust management to execute this pivot better than a buyer could. But that trust is fragile. We don't walk away from greed, we stay for trust, and in this case, the trust is based on a single quarter of earnings calls and a press release.
Takeaway: Actionable Price Levels The stock price of CORZ will now be driven by execution milestones, not narrative. The key levels to watch are the support at $12 (the price before the acquisition offer) and resistance at $18 (the offer price). If the company fails to deliver on the AMD partnership within the next two quarters, the stock will likely revert to the $12 level, and the shareholders who rejected the deal will have lost a guaranteed premium. Conversely, if Core Scientific announces a major customer for its AMD cluster, the stock could break above $18 and test the $25 level, which would imply a market cap above $9 billion. But I am not here to give price targets. Transparency is the shield against the next bubble, and right now, Core Scientific is not transparent enough. The company needs to publish monthly operational metrics: power consumption, GPU utilization, customer contracts, and software stack benchmarks. Until they do, the investment thesis is based on hope, not data. My advice to the community is simple: if you are invested in CORZ, demand transparency. Ask the management team to provide the same level of detail that a DeFi protocol would provide in a quarterly audit. If they refuse, treat the partnership as a marketing move, not a technical breakthrough. The market will eventually separate the hype from the substance, and when that correction comes, only the projects with real infrastructure will survive. Protect the flock, not just the profits, and that means being honest about the risks. The AMD partnership is a bet on a second-place chip maker. It is a bet that the AI infrastructure market will grow fast enough to absorb both Nvidia and AMD capacity. And it is a bet that Core Scientific's management can execute a complex engineering transition without the safety net of a cash exit. I have seen too many projects promise the moon and deliver a crater. This time, I am watching the data. And I suggest you do the same.