This text first appeared in The Vitality Magazine. The unique article could be considered right here. The Vitality Magazine (previously The Miner Magazine) gives information, knowledge, and insights on the power–compute–markets nexus.
Within the first installment of this sequence, we explored a foundational concept: Bitcoin mining was by no means nearly digital foreign money. It was designed as a long-term power system operating on a provide schedule that extends over greater than a century.
Within the second installment, we examined how that system will not be distinctive to Bitcoin. Fashionable AI knowledge facilities are constructed on the identical bodily basis—chips, energy, cooling, and infrastructure—all working collectively to show electrical energy into Bitcoin mining and AI processing at scale.
The AI increase has elevated demand exponentially, requiring cutting-edge infrastructure and high-efficiency expertise to assist grid resilience, in the end reshaping how the digital future is constructed.
Throughout the expertise sector, builders are more and more transferring past conventional grid reliance, securing electrical energy extra immediately via devoted power belongings. This “convey your personal energy” strategy might really feel new to AI, however it’s a playbook Bitcoin miners have been refining for years.
This type of vertical integration has grow to be a defining characteristic of the Bitcoin mining business. As corporations have undertaken enlargement into new markets, together with america, they’ve labored to establish current infrastructure. Nevertheless, this wave of enlargement has not stopped at knowledge facilities—it extends into energy itself, as operators search to enhance reliability, sustainability, and affordability of the power provide via direct possession and partnerships with power corporations.
This third installment builds on that evolution. If Bitcoin mining and AI knowledge facilities share the identical underlying system, the following query is how corporations place themselves inside it. What emerges is a spectrum of enterprise fashions—starting from outsourcing to full vertical integration.
The extra of this stack an organization controls, the extra capital-intensive it’s and the extra it influences its prices, efficiency, and in the end, its returns.
However it is very important word that this stack will not be static—and neither are the businesses working inside it.
As Bitcoin mining and AI knowledge facilities scale on the identical underlying infrastructure, the road between their enterprise fashions is starting to blur. The 2 seemingly distinct industries are more and more changing into one shared system, with stakeholders transferring throughout it in actual time.
The Entry Level: Asset-Mild Deployment
On the most elementary stage, participation in each Bitcoin mining and AI computing begins with deploying {hardware}.
In Bitcoin mining, this implies proudly owning ASIC machines—specialised chips designed to carry out a single process—and putting them into services operated by others to generate Bitcoin. In AI computing, the equal is deploying GPUs into knowledge facilities, the place they’re used to coach fashions and run workloads for patrons.
In each instances, corporations personal the machines—however not the infrastructure.
That infrastructure is offered by colocation operators, which provide energy, cooling, and bodily house to run compute at scale. Traditionally, this has been considered as a supporting perform. More and more, it’s changing into one of the crucial vital components of the enterprise.
Colocation is now not nearly internet hosting machines—it’s about monetizing energy and infrastructure.
Throughout Bitcoin mining, this mannequin has lengthy been embedded within the business. Firms like ABTC have deployed miners hosted via mum or dad infrastructure operators akin to Hut 8 (NASDAQ: HUT), whereas others, akin to Cango, function fleets hosted in services managed by Bitmain. In every case, the separation between {hardware} possession and infrastructure operation defines the enterprise mannequin.
That very same construction is now rising in AI.
Firms like Fluidstack are deploying GPU clusters via partnerships with infrastructure suppliers together with Cipher and TeraWulf, leveraging current power-connected websites to serve AI workloads with out proudly owning the underlying services. Bitdeer (NASDAQ: BTDR) is actively growing AI colocation capability at scale, together with a deliberate 180 MW facility in Tydal, Norway that’s presently in negotiations with a possible tenant.
As AI demand accelerates and energy turns into the limiting issue, infrastructure with current grid entry is changing into more and more helpful. Many websites initially constructed for Bitcoin mining are actually well-positioned to assist AI workloads, and over time, a rising share of those services—notably throughout the U.S. and Europe—are anticipated to transition towards AI and high-performance computing purposes.
In that sense, colocation is now not simply the entry level to the stack.
It’s changing into a bridge between two industries—connecting power, infrastructure, and compute demand in a single, evolving system.
Controlling Infrastructure
As corporations transfer up the stack, the following step is proudly owning the bodily setting itself.
At this stage, corporations are now not simply deploying {hardware}. As an alternative of counting on third-party internet hosting, operators construct or purchase their very own services, together with knowledge facilities, substations, and cooling methods.
This shift adjustments operations considerably. Infrastructure possession permits operators to manage energy prices, optimize efficiency, and scale back dependency on exterior suppliers.
However more and more, the worth of infrastructure is not only within the buildings—it’s within the energy connections hooked up to them.
That dynamic is now taking part in out throughout industrial belongings that had been as soon as thought-about out of date, permitting for corporations to show underutilized services into highly effective engines for development
Firms like Alcoa have begun exploring the sale of idled aluminum smelter websites to digital asset corporations akin to NYDIG, whereas Century Aluminum has bought a Kentucky facility to TeraWulf, which is now pivoting towards AI and high-performance computing.
Many of those websites confronted shutdown because of the gradual offshoring of high-paying business jobs to different nations. However they share one vital characteristic: they’re already linked to the power grid at scale.
That interconnection—typically the toughest and slowest a part of constructing new infrastructure—has all of the sudden grow to be a helpful asset in its personal proper.
In consequence, services initially constructed for heavy business are discovering a second life as compute infrastructure, repurposed to assist Bitcoin mining and AI workloads alike. That is bringing vital technical roles again to america, rebuilding key infrastructure and accelerating the nation as a worldwide chief in expertise and innovation.
On this setting, proudly owning infrastructure is now not nearly controlling operations. It’s about securing entry to power methods that may deal with growing demand whereas supporting total resilience.
Carry Your Personal Energy
However even that pool of grid-connected infrastructure is restricted.
The variety of industrial websites with current high-capacity interconnections is finite, and far of it has already been recognized or repurposed by main industries. As client demand for computing accelerates—notably from AI—the size of energy required necessitates options to take care of grid resilience whereas deploying new technological options.
In different phrases, the constraint is now not simply the place infrastructure exists. It’s whether or not the power grid itself can sustain. That strain is now forcing a broader shift.
Throughout main energy markets, operators are confronting a brand new actuality: connecting massive masses to the grid is changing into extra complicated and more and more unsure. Consequently, regulators are starting to revisit how massive power customers are built-in into the system.
In areas like PJM and ERCOT, grid operators have already began to regulate their frameworks in response to surging demand from knowledge facilities and different high-load customers. New guidelines and proposals are rising to manipulate how large-load knowledge facilities hook up with the grid, how prices are allotted, and the way reliability is maintained amid quickly rising demand.
To deal with these challenges, a rising variety of operators are transferring past the grid altogether.
A transparent instance of this strategy could be seen within the partnership between Amazon and Talen Vitality, the place knowledge middle infrastructure is being developed alongside nuclear era capability. Whereas AWS doesn’t personal the facility asset outright, the construction successfully aligns compute with a devoted power provide—mirroring the identical precept that has lengthy underpinned Bitcoin mining operations.
In Bitcoin mining, this has lengthy included co-locating with underutilized power sources. Firms like New West Knowledge flare fuel from oil manufacturing websites and use that energy to energise Bitcoin miners for additional money stream. In 2020, Greenidge Era turned what was referred to as the primary energy plant to immediately take part in bitcoin mining, reviving an asset that may have been shut down attributable to its lack of competitiveness within the energy market.
In AI computing, an identical mannequin is rising. Knowledge middle builders are more and more partnering with—or immediately constructing alongside—energy era belongings, together with pure fuel, nuclear, and most significantly, renewable power.
This “convey your personal energy” mannequin transforms electrical energy from a value middle right into a strategic benefit. It permits operators to stabilize pricing, guarantee availability, and align compute capability with power provide.
In Bitcoin mining, this evolution has been unfolding for years and is just rising.
A transparent instance is Bitfarms. The corporate traditionally operated as a self-mining enterprise, proudly owning infrastructure and deploying its personal computational energy. However with the acquisition of Stronghold, Bitfarms moved upstream into energy era, gaining direct management over power belongings. It later rebranded to Keel Infrastructure, signaling a broader transition away from pure Bitcoin mining towards a mannequin that may assist AI and high-performance computing workloads., and AI corporations are actually arriving on the identical conclusions. The important thing distinction is now not the system itself, however how every firm chooses to navigate.
Within the subsequent installment, we’ll take this one step additional: How these fashions are starting to converge—and what meaning for the way forward for power, compute, and capital.
This text first appeared in The Vitality Magazine. The unique article could be considered right here. The Vitality Magazine (previously The Miner Magazine) gives information, knowledge, and insights on the power–compute–markets nexus.
