We didn't.
Shareholders of Core Scientific (CORZ) voted down a $9 billion acquisition offer. Then, the company announced a partnership with AMD. The market cheered. But the silence in the ledger tells a different story. As a crypto media editor who has tracked mining infrastructure through two bear cycles, I've learned to read between the lines of press releases. The AMD deal is being hailed as a strategic pivot—but the data we need to validate it is conspicuously absent.
Context: The Ghost of Bankruptcy and the Pivot to AI
Core Scientific emerged from Chapter 11 bankruptcy in early 2024, carrying the scars of a brutal bear market. The company's core business remains Bitcoin mining, but the narrative has shifted to AI data center hosting—repurposing mothballed mining sites with liquid cooling and high-density racks to run GPU clusters for customers like CoreWeave. The $9 billion acquisition rejection by shareholders signals a belief that the company's intrinsic value exceeds that price. Now, with AMD as a chip supplier, the market is betting that the mining-to-AI conversion will unlock even more value.
But sentiment is a shifting tide, not a solid ground. The AMD partnership, announced with fanfare, lacks crucial details: no contract value, no minimum purchase commitments, no delivery timelines, no performance benchmarks. The press release reads like a strategic alignment, not a revenue contract.
Core: What the Press Release Didn't Say
Let me be blunt: I've spent years analyzing infrastructure plays—from the Raptor Protocol audit fiasco in 2018 to DeFi Summer's yield farming frenzy. The pattern is consistent. The market often treats a press release as a finished product, while the real work—engineering, deployment, optimization—lies ahead.
Core Scientific's advantage is its power infrastructure. The company holds long-term power purchase agreements at rates that would make any cloud provider jealous. That's the real asset. AMD's Instinct GPUs are a commodity; the power contracts are the moat. But converting a Bitcoin mine into an AI data center is not plug-and-play. It requires:
- Liquid cooling retrofits for dense GPU racks (mining rigs are air-cooled, A100s/H100s need direct-to-chip or immersion cooling).
- High-speed networking (InfiniBand or RoCE) to reduce GPU-to-GPU latency—a capability most mining sites lack.
- Software stack integration: AMD's ROCm is improving, but it still lags CUDA in library support and developer mindshare. Any AI customer will require proof that their workloads run efficiently on AMD hardware.
None of these technical hurdles are addressed in the announcement. The company has not disclosed how many megawatts (MW) of AI-ready capacity it has delivered, or the utilization rate of its existing GPU clusters. In a bear market, when survival matters more than gains, investors need to know if the company is burning cash to build capacity that may not find customers.
In the ledger’s silence, the true story whispers. The $9 billion rejection sets a valuation anchor, but it also creates a burden of proof. The market will now demand operational metrics to justify that valuation. The AMD partnership, without concrete numbers, is a narrative—not a catalyst.
Contrarian: The AMD Deal Could Be a Distraction
Here is the contrarian view that few are discussing: The AMD partnership might actually increase risk. By committing to a non-CUDA hardware stack, Core Scientific is betting on ROCm's maturity. If AMD's software ecosystem fails to gain traction, the company could be left with expensive GPUs that customers don't want. Meanwhile, competitors like CoreWeave are doubling down on Nvidia, whose CUDA ecosystem is the industry standard.
Furthermore, the rejection of the $9 billion acquisition means management must now deliver a business worth more than that. But the AMD partnership alone does not close the gap. The company needs to sign large, multi-year AI hosting contracts—not just chip supply agreements. The market is pricing in a future that may not materialize.
Every bull run is a myth waiting to be debunked. In a bear market, the debunking happens faster. The absence of a token means no liquidity mining or staking rewards to mask fundamental weaknesses. The stock price will reflect real operational performance, not speculative sentiment.
Takeaway: The Next 12 Months Will Reveal the Truth
Core Scientific has a unique asset: cheap power from mothballed mining sites. But that asset is only valuable if it can be converted into a revenue-generating AI infrastructure. The AMD partnership is a step toward supply chain diversification, but it is not a guarantee of customer demand. The company must now prove that it can deliver AI-ready capacity at scale, with competitive latencies and software compatibility.
In the next 12 months, the key metrics will be: megawatts of AI capacity deployed, utilization rates, and average contract length. If Core Scientific can demonstrate that its converted sites are running at 80%+ utilization with customers like CoreWeave, the stock will climb. If not, the $9 billion anchor will become a ceiling.
The narrative is shifting, but the tide has not turned. The true story is still being written in the silence of the ledger. We are watching.