From Crypto Mines to AI Gold: Core Scientific Powers AMD's Grand Strategy

MIAMI, FL – July 28, 2026 – In the relentless arms race for artificial intelligence supremacy, the battleground is quietly expanding. While headlines fixate on the processing power of next-generation chips, a far more fundamental war is being waged for the resources that make them run: power, land, and cooling. Advanced Micro Devices (AMD) just made a decisive move on this new front, announcing a strategic partnership with Core Scientific to secure up to 2.5 gigawatts (GW) of data center capacity. This isn't just a supply deal; it's a profound statement about the future of AI and a validation of a remarkable corporate rebirth.

At its core, the agreement involves AMD securing an initial 530 megawatts (MW) of high-density data center space from Core Scientific, set to come online starting in 2027. This initial phase alone represents over $14 billion in potential revenue for Core Scientific over its 15-year term. But the headline figure is the potential expansion to 2.5 GW—a staggering amount of power equivalent to what multiple nuclear reactors produce. For AMD, this move is about de-risking the deployment of its AI solutions for customers by ensuring the physical foundation exists to support them. For Core Scientific, it marks the crowning achievement of a high-stakes pivot from a bankrupt Bitcoin miner to an indispensable pillar of the AI ecosystem.

The New Currency of AI: Power and Real Estate

The AI boom, fueled by generative models and large-scale computing, has created an insatiable appetite for energy. According to industry analysts at Gartner, global data center electricity consumption is forecast to surge 26% year-over-year to 565 terawatt-hours in 2026. By 2027, AI-optimized servers are projected to consume more power than all conventional servers combined. This demand is creating a chokepoint far more restrictive than silicon fabrication capacity.

The challenge lies in power density. A traditional data center rack might draw 5-15 kilowatts (kW) of power. An AI rack packed with GPUs, however, can demand anywhere from 30 kW to over 100 kW. This intensity is overwhelming local power grids, with utility providers reporting that connecting new, large-scale data centers can take years longer than developers expect. Power availability has become the primary constraint on AI's growth.

This is the context that makes the AMD-Core Scientific deal so strategically vital. AMD isn't just buying space; it's securing a guaranteed “deployment path” for its Instinct™ GPUs and EPYC™ CPUs. “It's not about selling accelerators at this point, it's about selling a deployment path,” one principal analyst for data center technologies explained. By leveraging its balance sheet to lock in long-term capacity, AMD is removing the single biggest barrier for its customers—be they cloud providers or large enterprises—wanting to deploy its hardware at scale. The message is clear: the chip is only part of the solution. The real product is the entire, functioning, power-hungry AI cluster.

A Phoenix from the Crypto Ashes

Perhaps the most compelling narrative within this partnership is the transformation of Core Scientific itself. Just over two years ago, the company was emerging from Chapter 11 bankruptcy, a casualty of the crypto winter that decimated the Bitcoin mining industry. Today, it stands as a key enabler for the world’s most advanced technology sector. Its recent financial results tell the story of a stunningly successful pivot.

In the second quarter of 2026, Core Scientific’s total revenue more than doubled year-over-year to $164.2 million. The driver was its high-density colocation business, which generated $136.7 million—a staggering leap from just $10.6 million in the same period last year. Colocation for AI now accounts for 83% of its revenue, boasting a healthy gross margin of 59%. The company's legacy infrastructure, once used for energy-intensive Bitcoin mining, has proven to be its greatest asset. The large-scale facilities and, crucially, the massive power contracts it secured in its previous life are now AI gold.

This transition has not been without its challenges. The company reported a net loss of $1.15 billion in the same quarter, a figure driven primarily by non-cash accounting charges related to the rising value of its outstanding stock warrants as its share price soared. It has also undertaken massive capital expenditures—nearly $800 million in Q2 alone—to repurpose and build out its sites for AI clients. Yet, this investment is clearly paying off. Before the AMD deal, Core Scientific had already secured a landmark agreement with AI cloud provider CoreWeave for nearly 600 MW of capacity, a deal worth over $10 billion. The AMD partnership solidifies its position as a go-to provider for hyperscale AI infrastructure.

AMD's Full-Stack Gambit Against NVIDIA

For AMD, this partnership is a critical piece of its strategy to challenge NVIDIA’s dominance. The company is pursuing a full-stack approach, offering not just competitive hardware with its Instinct GPU series but also a robust software ecosystem with its ROCm™ platform. However, great technology is useless if customers can't deploy it. By collaborating with Core Scientific on the physical infrastructure design, AMD ensures its hardware and software will perform optimally in these massive, high-density environments.

“AI deployments are accelerating rapidly, and bringing that compute online requires trusted infrastructure partners with the scale and power to support the next era of AI,” said Mathew Hein, AMD’s senior vice president and chief strategy officer. This deal strengthens the entire AMD ecosystem, making it a more attractive and viable alternative for the model builders and cloud providers driving the AI revolution.

The inclusion of warrants in the deal—giving AMD the right to purchase up to 30 million shares of Core Scientific stock—is another shrewd move. It aligns the two companies’ long-term interests, giving AMD a direct stake in the success of the infrastructure it is helping to build. As Core Scientific's capacity and value grow, so too does the value of AMD's investment, creating a powerful symbiotic relationship.

The Gigawatt-Scale Future of Computing

The sheer scale of the AMD-Core Scientific agreement signals a new phase in the development of artificial intelligence. We are now in the era of gigawatt-scale computing, where success is measured not only in teraflops but in megawatts. The partnership is a template for the future, where technology companies must become deeply involved in the energy and real estate sectors to secure their roadmaps. This trend is already visible across the industry, with other former crypto miners like IREN and Hut 8 also leveraging their power-rich sites to land multi-billion-dollar AI cloud contracts.

With this deal, Core Scientific’s total leased customer power capacity now stands at approximately 1.1 GW, representing a pipeline of over $24 billion in contracted revenue. It is a testament to the idea that in a gold rush, selling picks and shovels—or in this case, powered real estate—is an immensely profitable business. As one Gartner analyst recently stated, “Building the compute capacity required for AI is the largest infrastructure project ever attempted by humanity.” In this monumental undertaking, companies that can deliver power at scale are no longer just service providers; they are the strategic gatekeepers to the future of technology.