I didn't see this coming. Shareholders of Core Scientific just rejected a $9 billion buyout. The stated reason? They believe an AMD partnership is worth more than that cash pile.
Alpha isn’t in the partnership announcement. It’s in what the market is ignoring: the gap between a press release and a working GPU cluster.
Context: The Infrastructure Play
Core Scientific is a listed company (CORZ). It’s a physical infrastructure play, not a smart contract protocol. Its core asset is not code or a treasury. It’s power contracts. Long-term, low-cost electricity agreements secured during the bear market of 2022-2023.
You don’t just mine Bitcoin with cheap power. You can also cool GPUs. This is the thesis: convert a Bitcoin mining facility into an AI data center. Rinse and repeat. The company has already signed multi-year hosting deals with CoreWeave. The AMD partnership is the next step in this “energy-to-compute” pivot.
But here’s the catch. The original article lacks any technical detail. No MW committed. No utilization rate. No testing results. Just a headline.
Core: The Order Flow Analysis
Let’s look at the order flow of capital. The $9 billion offer was a term sheet. By rejecting it, shareholders are effectively saying: “We can generate more than $9 billion in future value by ourselves.” This is a massive long-term bet on execution. The market doesn’t care about your vision. It cares about your P&L.
From a technical perspective, the AMD partnership is a risk, not a reward. I’ve built and deployed a trading bot on Ethereum L2s. I’ve felt the pain of unexpected failures. The AMD ROCm software stack is not as battle-tested as Nvidia’s CUDA. It’s a maturity gap. I’ve had to debug deployment scripts for ROCm. It’s not plug-and-play, especially for high-frequency inference workloads.
While the headlines screamed “Core Scientific partners with AMD!”, the reality is that converting a Bitcoin mine to a GPU cluster requires solving four new problems: liquid cooling, high-density rack power, InfiniBand networking, and GPU cluster orchestration. These are not trivial engineering challenges. They are multi-million dollar, multi-month integration projects.
And if AMD’s chip supply is constrained or yields are poor? The promised capacity won’t be delivered. The stock will get crushed.
Contrarian: The Retail vs. Smart Money Divergence
Retail sees the AMD partnership as a “Nvidia-antagonist” narrative. They think Core Scientific is “diversifying away from Nvidia risk.” Smart money sees the AMD partnership as a “power play” for AMD to get real-world testing data. AMD needs Core Scientific more than Core Scientific needs AMD. AMD needs to prove its Instinct GPUs can handle real AI workloads against Nvidia’s dominance. Core Scientific is essentially a guinea pig.
The real alpha isn’t in the AMD name. It’s in the power contract. The cheap electricity is the moat. Not the GPU. Not the partnership. The power contract is what allows Core Scientific to undercut traditional cloud providers on price. That’s the value proposition. The $9 billion offer was a reflection of that power contract value, not the AMD partnership.
So the contrarian view is: The AMD partnership is a distraction. It might actually dilute the company’s focus on its core power advantage. The market is ignoring the lack of revenue guarantees in the AMD deal. No minimum purchase. No revenue share. Just a hope.
Takeaway: The Forward-Looking Judgment
You don’t reject a $9 billion buyout for a partnership that hasn’t even delivered a single kW of compute. The market will punish this arrogance. The stock will likely trade sideways until the next quarterly report shows actual AI revenue.
If you’re long CORZ, you’re betting on one thing: execution. I’d rather be short the hype and wait for the first missed delivery date. The market doesn’t price in engineering delays. It only prices in the headlines. I’d wait for the real data: the settlement price of the first AMD GPU cluster, the utilization rate, and the power efficiency.
Alpha isn’t in the press release. It’s in the on-chain hash rate.