Hook

The market cheered. Core Scientific (CORZ) announced a 500 MW AI data center partnership with AMD, and the narrative shifted overnight: Bitcoin miners are reborn as AI landlords. But beneath the press release lies a detail the euphoria obscures—30 million shares of CORZ stock warrants granted to AMD. This isn't a simple supplier agreement. It's a structured financial instrument that redefines the risk calculus. Let me trace the wallet cluster of this deal: the capital flow, the dilution schedule, and the hidden puppeteer.

Context

Core Scientific, once the largest publicly traded Bitcoin miner by hashrate, emerged from Chapter 11 restructuring in January 2024. The company operates massive mining facilities primarily in Texas and North Carolina, with access to cheap power and hardened operational expertise. The post-bankruptcy strategy was clear: pivot to High-Performance Computing (HPC) and AI. AMD, the second-largest AI chip vendor after NVIDIA, needed a large-scale deployment partner to challenge NVIDIA's dominance alongside its own MI300X accelerators. The partnership announced on March 19, 2025, commits Core Scientific to host up to 500 MW of AMD-powered compute infrastructure. As part of the deal, AMD receives warrants to purchase up to 30 million shares of CORZ common stock. No exercise price was disclosed. This is the critical data point.

Core: On-Chain Evidence Chain of the Real Value Transfer

Let me be blunt: the 500 MW figure is impressive, but it's a capacity target, not a revenue guarantee. The real transfer of value is the 30 million warrants. Using my 2020 DeFi Liquidity Trap methodology, let's break down the economics.

First, the dilution risk. CORZ currently has approximately 180 million shares outstanding (post-restructuring). If the warrants are fully exercised at a hypothetical strike price of, say, $5 (near current trading range), AMD would inject $150 million into Core Scientific. That cash is a lifeline for the $2-3 billion capital expenditure required to build 500 MW of AI-ready infrastructure. But existing shareholders would face a 16.7% dilution. In a bull market, dilution is often ignored. But every data detective knows: liquidity is not value; flow is the truth. The flow here is new shares entering the market, potentially overhanging the stock.

Second, the incentive structure. AMD is not just a vendor; it's now a large insider. This mirrors the "strategic investor" pattern I observed in 2017 ICOs where foundations gave tokens to exchanges for listings. Here, AMD gets a call option on Core Scientific's success. If CORZ stock soars to $50, AMD's warrants become worth $1.35 billion. That is a massive incentive for AMD to ensure Core Scientific succeeds—by providing preferential chip pricing, technical support, and perhaps even customer leads. But it also means AMD's interests are aligned only as long as the stock price appreciates. Should CORZ falter, AMD can simply let the warrants expire. Whales do not whisper; they dump on the charts.

Third, the 500 MW timeline. Building a hyperscale data center takes 18-36 months. The first 100 MW may come online by mid-2026. In the interim, Core Scientific must continue its Bitcoin mining operations, which are exposed to BTC price volatility. The transition is not instantaneous. Using my 2022 Terra post-mortem framework, I see a classic commitment trap: a company borrows against future promise to fund present transformation, creating a cash-flow bridge that can collapse if the AI narrative fades or chip delivery delays occur.

From a competitive positioning perspective, Core Scientific chose AMD over NVIDIA. This is a bet that AMD's MI300X will capture meaningful market share from NVIDIA's H100/B200. Smart contracts execute; humans manipulate. The manipulation here is AMD's market positioning—using Core Scientific as a showcase customer to lure other miners and enterprises. But the wallet cluster of AI compute demand still heavily favors NVIDIA. Major AI labs (OpenAI, Anthropic, Google) are locked into NVIDIA's CUDA ecosystem. Core Scientific is effectively building a data center to host AMD chips for a secondary market. That is a higher-risk, potentially higher-reward strategy.

Contrarian: Correlation Is Not Causation—The Five Hidden Blind Spots

The market is treating this deal as a pure positive catalyst. I see five structural blind spots.

1. Capital Expenditure Overhang – Building 500 MW requires $2-3 billion. Core Scientific had ~$80 million in cash post-restructuring. The AMD warrant cash ($150 million at best) is a fraction. Where does the rest come from? Debt or secondary equity offerings. Dilution will be larger than the stated 30 million shares. My 2020 experience warning about hidden leverage in DeFi applies here: off-balance-sheet commitments are the new protocol risk.

1. Counterparty Risk Concentration – The entire deal depends on AMD delivering MI300X chips on schedule and at volume. AMD's supply chain is not infinite. If NVIDIA launches a more competitive product, AMD's deployment could be delayed or reduced. Core Scientific has no backup plan publicly disclosed.

1. Power Arbitration Risk – Core Scientific's advantage is cheap power from Bitcoin mining. But AI data centers require firm, uninterruptible power, not the interruptible load agreements typical for miners. Retrofitting facilities for liquid cooling and high-density power distribution is capital-intensive. The cost per MW for AI is 3-5x a standard mining container.

1. Customer Concentration – Who will fill 500 MW of capacity? Core Scientific has not named a single anchor tenant. In the AI cloud market, hyperscalers (AWS, Azure, GCP) dominate. Core Scientific is competing against CoreWeave (backed by NVIDIA, valued at $19B) and established colocation providers. Without a marquee client, the data center is a spec build.

1. The Narrative Trap – Every cycle, miners pivot to a hot new narrative to justify equity value. In 2021 it was "green mining" and "carbon credits." In 2023 it was "mining-as-a-service." Now it's AI. The success of these pivots has been mixed. Due diligence is the only hedge against hype.

Takeaway: The Next-Week Signal

The true test of this deal's merit will come not from the announcement but from the SEC filing that discloses the warrant exercise price. If the strike is below $4 (close to CORZ's post-restructuring floor), AMD is getting a bargain, and the deal signals short-term intent rather than long-term alignment. If the strike is above $8, it's a confident bet. Track the filing. Also monitor Core Scientific's monthly operational updates for any mention of LOIs with AI tenants. Until then, the data shows a transformation in progress, but not yet a transformation earned. The balance sheet tells the story before the stock market does.

Tracing the seed round to the exit strategy — the warrants are the seed. Liquidity is not value; flow is the truth — follow the cash from the warrants to the construction contracts. Whales do not whisper; they dump on the charts — watch for insider selling post-announcement.