We didn't think they'd say no. Nine billion dollars. A clean exit. A chance to walk away from the noise of Bitcoin mining and the weight of a 2023 bankruptcy. But Core Scientific's shareholders did exactly that. They rejected the acquisition offer, and then, almost as if on cue, the company announced a partnership with AMD. The market cheered. The stock jumped. And I sat in my Istanbul apartment, staring at the press release, feeling the familiar itch of skepticism.
I've been in this industry long enough to know that the loudest announcements often hide the quietest doubts. From the DeFi Summer of 2020 to the NFT mania of 2021, I've watched companies ride waves of hype only to crash when the technical reality caught up. Core Scientific's story is different—it's a physical infrastructure play, not a smart contract. But the same principles apply: without verifiable data, every press release is just a story. And stories, as we learned in the bear market, can be fragile.
So let's dig into what this $9 billion 'no' really means, and whether the AMD partnership is the lifeline it appears to be, or just another chapter in the long, messy history of mining companies trying to reinvent themselves.
Context: From Mining to AI Hosting
Core Scientific is not a new name in crypto. Before the 2022 crash, it was one of the largest publicly traded Bitcoin miners in North America, with massive facilities in Texas and other states. The company filed for Chapter 11 bankruptcy in December 2022, crushed by falling Bitcoin prices, rising energy costs, and a debt load that had become unsustainable. It emerged from bankruptcy in January 2024, having restructured its balance sheet and secured new financing.
But the company didn't just return to mining. It pivoted. The thesis was simple: the same infrastructure that powers Bitcoin mining—high-capacity power, cooling systems, and secure facilities—can be repurposed for AI data centers. AI workloads, particularly training and inference for large language models, require enormous amounts of electricity and compute. The bottleneck isn't just GPUs; it's the physical infrastructure to host them. Core Scientific saw an opportunity to become a hosting provider for AI compute, leveraging its existing sites and power purchase agreements.
The pivot began before the bankruptcy. In 2023, Core Scientific signed a multi-year hosting contract with CoreWeave, a specialized AI cloud provider. That deal was the first real validation of the transition. But the AMD partnership, announced in early 2024 alongside the rejection of the $9 billion acquisition, took things a step further. AMD is not just a chip supplier; it's a strategic partner in the AI arms race, desperate to compete with Nvidia's dominance.
We didn't have the technical details when the news broke. No hashrate, no megawatts, no GPU count. Just a press release that sent the stock up. But for those of us who remember the DeFi Summer hype cycles, the pattern is familiar. The market loves narratives. And the narrative of 'AI + crypto' is irresistible. But narratives are not data.
Core: The Technical Reality of the AMD Partnership
Let's start with what we actually know. The partnership was announced, but the terms were not disclosed. No contract value, no minimum purchase commitments, no timeline for deployment. The only concrete detail is that Core Scientific will use AMD's Instinct GPUs to power its AI cloud services. That's thin. Very thin.
I've spent years auditing infrastructure projects, from DeFi protocols to Layer 2 scaling solutions. The one thing I've learned is that when a company announces a partnership without numbers, it's usually because the numbers are too small or too uncertain to share. The AMD partnership could be a pilot program involving a few hundred GPUs, or it could be a multi-year commitment worth hundreds of millions. We don't know. And that uncertainty is a red flag.
From a technical perspective, converting a Bitcoin mining facility to an AI data center is not trivial. Mining rigs are ASICs—specialized chips designed for a single task: hashing. They produce heat, but they are relatively simple to cool. GPU clusters, on the other hand, require sophisticated liquid cooling, high-density networking (InfiniBand or RoCE), and a software stack that can orchestrate distributed training. The infrastructure is fundamentally different.
I've walked through mining facilities in Istanbul. They are loud, hot, and designed for ASICs. Converting them to hold GPU clusters is like turning a cargo ship into a cruise liner. Possible, but not without a complete overhaul. The power capacity might be there, but the cooling, the networking, the layout—all of it needs to be rethought. Core Scientific has experience with hosting, thanks to the CoreWeave deal, but scaling that to a new partnership with AMD introduces new challenges.
The biggest challenge is the software stack. AMD's GPUs use the ROCm open-source platform, while Nvidia's CUDA is the industry standard. ROCm has improved significantly, but it still lags behind CUDA in terms of library support, developer tools, and ecosystem maturity. Any AI company that wants to use AMD GPUs will have to port its code to ROCm, which is a non-trivial engineering effort. Core Scientific, as a hosting provider, is not responsible for that porting—that's the client's job. But the availability of AMD GPUs alone doesn't create demand. Clients need to see a clear path to running their workloads on AMD hardware. If the ecosystem isn't ready, the GPUs will sit idle.
We didn't see any mention of this in the press release. No announcement of a software partnership, no reference to ROCm optimization, no case studies of existing AMD-based AI deployments. The silence is telling.
The $9 Billion Anchor
Now, let's talk about the elephant in the room: the rejected acquisition. The offer was $9 billion, which valued Core Scientific at a significant premium to its pre-announcement market cap. The shareholders said no, betting that the company could create more value through the AMD partnership and its AI pivot. That's a bold bet. It means the board and major shareholders believe Core Scientific is worth more than $9 billion in the long run.
But what does that valuation depend on? It depends on the success of the AI hosting business. It depends on the AMD partnership delivering real revenue. It depends on the company maintaining its mining operations as a cash cow while the AI business scales. That's a lot of dependencies.
I've seen this pattern before. In the crypto world, companies often reject acquisition offers because they believe in their own vision. But visions don't always materialize. The market is now pricing in a successful pivot. If the AMD partnership fails to deliver, the stock will fall, and the $9 billion offer will look like a missed opportunity. The pressure is on.
Contrarian: The Unspoken Risks
Here's the contrarian take: The AMD partnership might be more valuable for AMD than for Core Scientific. AMD needs real-world deployment to validate its GPUs. It needs reference architectures, case studies, and happy customers. Core Scientific provides the lab. But the relationship is asymmetric. If AMD's chips underperform, Core Scientific's clients could leave. If AMD's supply chain falters, Core Scientific's expansion stalls. The market is pricing in a best-case scenario that ignores the operational risks.
There's also the debt hangover. Core Scientific emerged from bankruptcy with a cleaner balance sheet, but it still carries debt. The AI pivot requires capital expenditure—new cooling systems, networking equipment, and the GPUs themselves. If the company needs to raise more capital, it could dilute existing shareholders. The $9 billion offer might have been a lifeline, but the shareholders chose the riskier path.
And let's not forget the Bitcoin mining side. The company still has a large mining operation. The 2024 halving cut block rewards in half, squeezing margins. Core Scientific has been successful in securing low-cost power, but the mining revenue is now more volatile than ever. The AI business is supposed to diversify revenue, but it also adds complexity. Managing two very different businesses under one roof is a challenge that many companies have failed at.
We didn't hear much about the mining operations in the press release. It was all about AI. But the mining side is still the foundation. If Bitcoin prices fall, the mining cash flow could dry up, leaving the AI business to carry the weight. That's a dangerous dependency.
Takeaway: The Truth Is in the Data
So where does this leave us? Core Scientific's $9 billion 'no' is a statement of intent. The company is betting on the AI pivot, and the AMD partnership is a key piece of that bet. But the evidence is thin. The press release lacks the technical details that would give us confidence. The ecosystem risks are real. The valuation expectations are high.
In my experience, the best way to evaluate a company like Core Scientific is to ignore the press releases and focus on the operational metrics. How many megawatts of AI capacity have they deployed? What is the utilization rate? What is the revenue per megawatt? Are they retaining clients? These are the numbers that matter. Until we see them, the AMD partnership is just a story.
We didn't build blockchain to trust press releases. We built it to verify data. Apply the same skepticism here. The market is euphoric about AI-crypto convergence, but euphoria is a poor guide to reality. The real test isn't the announcement. It's the quarterly earnings report. That's where the truth lives.
I'll be watching. And I'll be writing. Because in a bull market, the most valuable thing is a clear-eyed view of the risks. And right now, Core Scientific's AMD bet is a high-risk, high-reward gamble that could go either way. The $9 billion 'no' was a vote of confidence. Now we need to see if that confidence was justified.
As I sit here in Istanbul, watching the Bosphorus flow, I'm reminded of a lesson I learned during the 2022 crash: narratives are cheap. Data is expensive. And the companies that survive are the ones that deliver data, not just stories. Core Scientific has a chance to be one of those companies. But the AMD partnership is not a silver bullet. It's a test. And the results are not yet in.
Let's keep our eyes on the numbers. Because that's where the real story is.
— Chloe Martin, Istanbul, 2024
