I didn’t see this coming. But then again, I’ve been in this game long enough to know that when shareholders reject a $9 billion exit, they’re not playing defense. They’re playing offense.
Core Scientific, the Nasdaq-listed bitcoin mining and AI data center operator, just told the market that its future is worth more than $9 billion. The offer was on the table. The Board said yes. But the shareholders—the people who actually own the company—said no. And they did it while simultaneously announcing a chip partnership with AMD that’s supposed to be the real prize.
Let me rewind for you. This is a company that survived bankruptcy in 2023. It emerged leaner, meaner, and with a new narrative: "We’re not just miners anymore. We’re AI infrastructure." And now, they’re betting the house on that pivot.
Context: Why Now? Core Scientific is a hybrid beast. It started as a pure-play bitcoin miner, building massive data centers in Texas and other energy-rich locations. But after the 2022 crypto winter and the bankruptcy filing, management realized that mining alone is a commodity business with razor-thin margins. The real money? HPC (high-performance computing) and AI cloud services. Think CoreWeave, but with legacy mining assets.
In 2024, they signed a multi-year hosting deal with CoreWeave. That was the first signal. Now, the AMD partnership is the second. But here’s the kicker: the shareholders just rejected a $9 billion acquisition offer. That means they believe the company, post-AMD, is worth more than that. And they’re willing to wait for it.
Core: The AMD Partnership – What We Actually Know According to the announcement, Core Scientific will deploy AMD’s Instinct GPUs in its data centers. That’s it. No specific wattage, no number of GPUs, no timeline for delivery. Just a press release and a partnership handshake.
Now, let’s be real. This is not a technological breakthrough. This is a supply chain diversification move. AMD has been struggling to break Nvidia’s stranglehold on the AI chip market. Core Scientific, with its existing infrastructure and power contracts, becomes a real-world testing ground for AMD’s ROCm software ecosystem. It’s a win-win: AMD gets a customer to prove its hardware, Core Scientific gets a cheaper alternative to Nvidia’s H100s.
But here’s the part that the market is missing. Converting a bitcoin mining facility to an AI data center is not plug-and-play. Bitcoin miners use ASICs (Application-Specific Integrated Circuits) that run on air cooling. AI GPUs need liquid cooling, high-density racking, InfiniBand networking, and a completely different power distribution setup. Core Scientific has the real estate and the power purchase agreements (PPAs) that lock in cheap electricity—that’s the real asset. But the capex for retrofitting is massive. The article didn’t disclose how much they’re spending on this conversion. That’s a red flag.
Contrarian: The Shareholder Vote Is the Real Story, Not the AMD Deal Every headline is screaming about the AMD partnership. But the most important signal is the rejection of the $9 billion sale. Think about it: the company was in bankruptcy just 18 months ago. Now, shareholders are saying, "We’re worth more than nine figures." That’s either incredible confidence or dangerous hubris.
I’ve been in the room when deals like this fall apart. In 2017, I watched a small Canadian exchange reject a $50 million buyout offer because they thought they could grow faster. Six months later, they were delisted. The difference? Core Scientific has a real asset: power. And power is the new oil.
But here’s the contrarian take: the AMD partnership might actually be a distraction. The real value driver for Core Scientific is not the chip brand; it’s the ability to deliver megawatts of compute. The market is focusing on the headline, but the execution risk is enormous. Every day that AMD doesn’t ship, the narrative shifts. Algorithms smell fear, but they respect speed. Right now, Core Scientific is betting on AMD’s speed. And AMD has a history of missing deadlines.
My Take: The Human Side of the Bet I’ve seen this movie before. During the Terra/Luna collapse, I hosted a roundtable in Toronto. Traders were crying into their coffee. The ones who survived were the ones who understood that narrative velocity matters more than fundamentals—until it doesn’t.
Core Scientific’s shareholders just raised the stakes. They said no to $9 billion because they believe the company can create more value by executing the AI pivot. That’s a bet on the management team, on the power contracts, and on AMD’s ability to deliver. But it’s also a bet on the market’s willingness to pay for a story.
Yield is a drug; exit liquidity is the cure. In this case, the shareholders chose the drug. They want the long-term high of AI growth, not the short-term fix of a sale.
What to Watch Next Forget the AMD partnership memes. The real metric to watch is the number of megawatts (MW) delivered to AI customers. If Core Scientific can convert 100 MW of its mining capacity to AI hosting by Q3 2025, the stock will fly. If they stumble on the engineering—cooling, networking, ROCm stability—the narrative will crack.
Chaos is just data waiting for a narrative. Right now, the narrative is bullish. But the data is thin. We don’t know the capex, we don’t know the timeline, and we don’t know if AMD’s GPUs can actually compete with Nvidia in production.
I’ll be watching the earnings calls. The real story is in the numbers, not the press releases. And if the shareholders are right, this could be one of the biggest turnaround stories of the decade. If they’re wrong, it’s a cautionary tale about hubris.
Either way, it’s a hell of a story. And I’m here for it.
Final Thought We don’t bet against people who have survived bankruptcy. Core Scientific burned to the ground and rebuilt. That takes grit. But grit doesn’t pay the bills. Execution does. The next 12 months will tell us if this dare was a smart move or a tragic mistake.
For now, I’m holding my breath. And I’m watching the power meters.