Bitcoin Mining and AI: The Infrastructure Convergence

The Great Pivot is underway. Across North America, the largest publicly traded Bitcoin mining companies are systematically converting their facilities into AI data centers. Hut 8 signs a $7 billion deal backed by Google. IREN (formerly Iris Energy) locks in a $9.7 billion contract with Microsoft. Core Scientific becomes the backbone of CoreWeave’s AI empire. The numbers are staggering, the headlines relentless, and the narrative is clear: Bitcoin mining infrastructure is being absorbed by the artificial intelligence industry.

But here is the part most analysts miss—and the part that matters most to the Mining Hackers community: this is actually good news for home miners.

As industrial miners chase AI dollars, they are vacating Bitcoin’s hashrate battlefield. That creates more room for pleb miners, improves the economics of solo mining, and makes the decentralization mission more critical than ever. At D-Central Technologies, we are not pivoting anywhere. We are doubling down on what we have always done: hacking institutional-grade mining technology into accessible solutions for the sovereign individual.

Let us break down what is happening, why it matters, and what it means for everyone running a miner at home.

The Convergence: Why Bitcoin Mining Infrastructure and AI Compute Are Merging

The overlap between Bitcoin mining operations and AI data centers is not coincidental—it is structural. Both industries require the same three things in massive quantities: cheap electricity, industrial-scale cooling, and real estate with robust grid connections.

A Bitcoin mining facility consuming 200 MW of power with liquid cooling infrastructure, high-voltage substations, and fiber connectivity is, from an infrastructure standpoint, almost identical to what a hyperscale AI training cluster requires. The hardware inside is different—ASIC miners versus NVIDIA GPUs—but the shell, the power delivery, and the thermal management are the same.

This convergence was inevitable for several reasons:

• Power contracts are the bottleneck. AI companies need megawatts of cheap power yesterday. Bitcoin miners have already spent years negotiating favorable long-term power purchase agreements. Acquiring a mining facility is faster than building from scratch.
• Cooling expertise translates directly. Mining operations have pioneered immersion cooling, hot/cold aisle containment, and desert-climate thermal management. These are exactly the capabilities AI data centers need for GPU-dense environments.
• Grid relationships are worth billions. The interconnection agreements, utility partnerships, and grid stability contracts that miners hold can take years to establish. AI companies would rather buy them than wait.
• Rural data center experience. Many mining operations are located in low-cost, remote areas near hydroelectric or natural gas sources. AI companies are running out of space in traditional data center markets like Northern Virginia and are now looking at exactly these locations.

The result is a $65 billion wave of AI contracts announced by Bitcoin miners through late 2025, with GPU data centers generating three times the revenue per megawatt compared to traditional Bitcoin mining operations. HIVE Digital Technologies estimates that 10 MW of NVIDIA H100 GPUs can generate revenue equivalent to 100 MW of Bitcoin mining.

Who Is Pivoting: The Biggest Names Making the Switch

The list of companies redirecting their Bitcoin mining infrastructure toward AI reads like a who’s who of publicly traded miners:

IREN (Formerly Iris Energy)

IREN has arguably made the most dramatic transformation. The company secured a landmark $9.7 billion, five-year AI cloud agreement with Microsoft in late 2025, cementing its transition from pure-play Bitcoin miner to AI cloud infrastructure leader. With a $14 billion market capitalization and a 300% year-to-date stock gain in 2025, IREN is the poster child for the pivot. The company still mines Bitcoin but its identity and valuation are now driven entirely by AI compute.

Hut 8

The Canadian-founded company made headlines with a $7 billion, 15-year AI data center lease for 245 MW at its River Bend site in Louisiana, backed by a Google-affiliated financial framework. Hut 8’s stock surged approximately 139% on the news, with an 80% rally throughout 2025. What was once a Canadian Bitcoin mining champion is now primarily an AI infrastructure company operating south of the border.

Core Scientific

Perhaps the most dramatic turnaround story in mining history. After emerging from bankruptcy in 2024, Core Scientific signed hosting deals with CoreWeave that expanded to approximately $10.2 billion in projected cumulative revenue over 12-year terms, covering roughly 590 MW across six sites. The company was so valuable as AI infrastructure that CoreWeave offered to acquire it for approximately $9 billion—an offer that Core Scientific initially rejected. A $1.2 billion expansion at its Denton, Texas site followed. The former Bitcoin miner is now the dominant colocation provider for AI workloads.

Bitdeer

Bitdeer has taken a dual approach, operating AI cloud services on NVIDIA DGX SuperPOD platforms while simultaneously developing its own custom ASIC chips for Bitcoin mining. The company reports near-full utilization of its AI infrastructure—a hedge-everything strategy that reflects the uncertainty of this transition period.

Marathon Digital

With approximately 40 EH/s of Bitcoin mining hashrate and over 600 MW of data center capacity across North Dakota, Texas, and the UAE, Marathon is the largest remaining Bitcoin-focused miner among publicly traded companies—though even Marathon has publicly moved into AI computing, unable to ignore the revenue differential.

HIVE Digital Technologies

Operating facilities in Sweden, Canada, and Iceland—all powered by renewable energy—HIVE was among the first to articulate the economics clearly: 10 MW of NVIDIA H100 GPUs generates revenue comparable to 100 MW of Bitcoin mining. That 10x revenue multiplier per megawatt is the single number driving the entire industry pivot.

Why Bitcoin Miners Make Natural AI Hosts

It is important to understand why AI companies are specifically targeting Bitcoin miners rather than building their own infrastructure from the ground up:

1. Pre-negotiated cheap power. Bitcoin miners have spent years securing some of the lowest electricity rates in North America. Power costs represent 30-40% of AI data center operating expenses, making these contracts extraordinarily valuable.
2. Proven cooling at scale. Mining facilities have deployed and iterated on thermal management systems for years. AI GPU clusters like the NVIDIA H100 and the upcoming B200 generate extreme heat density that requires exactly this expertise.
3. Speed to market. Building a new data center takes 18-36 months. Converting a mining facility can be done in 6-12 months. In the AI arms race, time is the scarcest resource.
4. Grid interconnection. The electrical interconnection to the grid—substations, transformers, transmission access—is often the longest lead-time item for a new data center. Miners already have this in place.
5. Regulatory and permitting experience. Miners have already navigated the local zoning, environmental permitting, and community relations required to operate power-intensive facilities. This institutional knowledge is difficult to replicate.
6. Flexible demand profile. Bitcoin miners are accustomed to curtailing operations during grid stress events, which makes them attractive partners for utilities. AI companies benefit from inheriting these grid-friendly relationships.

The Bull Case: Why Wall Street Is Excited

From a pure financial perspective, the numbers are compelling. Companies that have successfully pivoted are reporting 80-90% operating margins on AI hosting contracts, compared to the razor-thin margins many miners experienced after the April 2024 halving. AI contracts generate three times the revenue per megawatt. Stock prices have reflected this: IREN up 300%, Hut 8 up 139%, the entire sector re-rated from “volatile crypto play” to “AI infrastructure.”

The bull case is straightforward:

• Diversified revenue streams protect against Bitcoin price and difficulty volatility.
• Higher margins on AI compute versus Bitcoin mining, especially post-halving.
• Infrastructure value unlock—the power contracts and grid connections are worth more than the mining operations they support.
• Institutional capital access—AI infrastructure attracts investment from hyperscalers like Microsoft and Google that would never invest in Bitcoin mining.
• Long-term contract visibility—12 to 15-year hosting deals provide revenue predictability that Bitcoin mining never offered.

J.P. Morgan, Goldman Sachs, and other major banks have published research notes re-categorizing former Bitcoin miners as AI infrastructure plays, complete with higher price targets and upgraded ratings.

The Bear Case: What the Financial Analysts Are Not Telling You

Now for the part that matters to the cypherpunk community. The financial press is treating this pivot as pure upside. But there are serious risks and ideological concerns that deserve honest examination.

The Technical Risks

• ASICs are not GPUs. Bitcoin mining hardware is useless for AI computation. Miners must acquire or lease NVIDIA H100s, A100s, or MI300X GPUs plus high-speed interconnects—capital expenditures of hundreds of millions of dollars. A mining facility cannot simply be “switched” to AI.
• Uptime requirements are different. Bitcoin miners can tolerate brief outages without catastrophic consequences. AI training runs can be ruined by a single interruption. The reliability standards for AI hosting are significantly more demanding.
• Talent gaps are real. Teams trained on managing ASIC miners and cryptocurrency pools must rapidly upskill to manage GPU clusters, machine learning frameworks, and enterprise SLAs. These are fundamentally different skill sets.
• Counterparty risk. A 12-year contract with a single AI customer like CoreWeave creates concentration risk. If the AI boom cools or a customer defaults, the miner is left with expensive GPU infrastructure and no revenue.
• The AI bubble question. The entire pivot thesis depends on sustained, massive demand for AI compute. If AI training efficiency improves dramatically (as DeepSeek has demonstrated), or if the current investment cycle corrects, miners who sold their Bitcoin infrastructure may have nothing to fall back on.

The Ideological Concerns

This is where D-Central’s perspective diverges sharply from Wall Street’s.

These companies built their mining operations on the narrative of decentralization, sound money, and securing the Bitcoin network. Their investors, customers, and communities supported them because they believed in Bitcoin’s mission. Now, at the first sign of more profitable pastures, they are abandoning that mission to serve NVIDIA, Microsoft, and Google—the very centralized technology giants that the cypherpunk movement was designed to counterbalance.

Consider the irony: infrastructure that was built to decentralize money is now being repurposed to centralize AI power in the hands of a few trillion-dollar corporations. The same megawatts that once secured the most important open, permissionless financial network in human history are now training proprietary AI models owned by companies that would not hesitate to censor, surveil, or gatekeep.

This is not a neutral business decision. It is a philosophical retreat.

The Impact on Home Miners: Why the AI Pivot Is Actually Good News

Here is the silver lining that nobody in traditional finance is talking about, because they do not care about home miners. But we do.

When large industrial miners divert hundreds of megawatts from Bitcoin mining to AI hosting, they are removing hashrate from the Bitcoin network. Every exahash that goes offline to make room for GPU racks is hashrate that is no longer competing with your miner at home.

What does this mean in practice?

• Difficulty adjustments become more favorable. If significant industrial hashrate leaves the network, Bitcoin’s difficulty algorithm adjusts downward every 2,016 blocks, making it easier for remaining miners—including home miners—to find blocks.
• Solo mining odds improve. If you are running a Bitaxe or other open-source solo miner, your probability of hitting a block increases as total network hashrate decreases. The lottery gets slightly better odds.
• Hashprice recovers. In 2025, hashprice dropped to historic lows of approximately $35 per petahash per second, with average costs reaching $70,000 per mined Bitcoin. If hashrate growth slows due to the AI diversion, hashprice stabilizes or improves for those who remain.
• Pool mining payouts increase. For home miners using pools, fewer total network participants means your share of each block reward grows proportionally.
• Used equipment becomes available. As industrial miners offload their ASIC inventory to make room for GPUs, the secondary market gets flooded with used mining hardware at lower prices. This is great news for anyone looking to start Bitcoin mining or expand their home operation.

In the simplest terms: the big players leaving for AI means more room for the little guys. The playing field is leveling, and that is exactly what Bitcoin was designed to enable.

What This Means for D-Central’s Community

D-Central Technologies exists to serve the home miner, the solo miner, the pleb who believes that decentralizing Bitcoin’s hashrate is not a business strategy but a moral imperative. Our position on the AI pivot is straightforward:

We acknowledge the trend honestly. The economics driving industrial miners toward AI are real. A 10x revenue multiplier per megawatt is not something any publicly traded company’s board of directors can ignore. We are not blind to why this is happening.

But we are not pivoting. D-Central was built for a different mission. We are Mining Hackers—taking institutional-grade technology and making it accessible for the sovereign individual. We build Bitaxe accessories, we repair ASICs, we design space heaters that mine Bitcoin while they warm your home, and we educate the community on how to mine effectively from anywhere. None of that changes because Wall Street found a new narrative.

What changes for our community is this:

• Open-source mining hardware matters more. As corporate miners exit Bitcoin for AI, devices like the Bitaxe and NerdAxe become more important for network security. Every Bitaxe plugged into a home outlet is a small act of decentralization that keeps the network resilient.
• Dual-purpose mining becomes more compelling. Bitcoin space heaters that monetize heat while mining are a uniquely home-miner solution. As industrial miners leave, the economic case for mining-while-heating only gets better. Our product lineup is designed precisely for this use case.
• Education is critical. New miners entering the space need to understand the equipment, the economics, and the mission. D-Central’s educational content, repair guides, and community support are more valuable now than ever.
• Canadian mining has an edge. Bitcoin mining in Canada benefits from cold climate cooling advantages, relatively affordable energy, and regulatory stability. As US-based industrial miners chase AI contracts with American hyperscalers, Canadian home miners remain focused on Bitcoin.

The Decentralization Angle: Why Home Miners Are Now More Important Than Ever

This is the most important section of this article, and it is the perspective you will not find in any financial analyst’s research note.

Bitcoin’s security model depends on a distributed network of miners. The more geographically dispersed and independently operated the hashrate is, the more resistant the network is to censorship, state-level attacks, and regulatory capture. This is not abstract philosophy—it is Bitcoin’s core security architecture.

In January 2025, a severe winter storm in the United States caused Foundry USA Pool to lose approximately 200 EH/s of hashrate—roughly 60% of its capacity went offline in a single weather event. That is what happens when hashrate is concentrated in large industrial facilities in a single geographic region. Centralized mining is fragile mining.

Now consider what happens as industrial miners exit Bitcoin for AI:

• The percentage of hashrate from home miners increases. Even if the absolute number of home miners stays the same, their proportional contribution to network security grows as industrial miners leave.
• Geographic distribution improves. Home miners are spread across hundreds of thousands of locations worldwide. No single weather event, regulatory action, or grid failure can take them all offline simultaneously.
• Censorship resistance strengthens. A network secured by millions of independent home miners is fundamentally harder to censor than one secured by a handful of corporations who are now beholden to Microsoft and Google for their AI revenue.
• The incentive alignment improves. Home miners who mine Bitcoin are Bitcoiners. They are aligned with the network’s long-term health. Industrial miners who were always just chasing the best return on capital have proven their loyalties lie elsewhere.

The United States currently holds 37.5% of global Bitcoin hashrate at approximately 400 EH/s. If even a fraction of that industrial hashrate migrates to AI workloads, the relative importance of distributed home mining increases proportionally. Every Bitaxe, every NerdAxe, every space heater mining Bitcoin in someone’s garage becomes a slightly more important node in the network’s security architecture.

This is the real story of the AI pivot: not that Bitcoin mining is dying, but that Bitcoin mining is being purified. The mercenary capital is leaving. What remains are the true believers—the miners who are here for the mission, not the margins.

Looking Forward: The Mining Hackers Stay the Course

The next two to three years will reshape the Bitcoin mining industry more dramatically than any halving event. Mining revenue as a percentage of total revenue for publicly traded miners is projected to fall from 85% in early 2025 to under 20% by late 2026 for companies that have committed to AI contracts. The companies that built their brands on Bitcoin will increasingly be AI infrastructure companies that happen to still mine some Bitcoin on the side.

For the Mining Hackers community—for the thousands of home miners running Bitaxes, NerdAxes, space heaters, and repurposed Antminers—this transition represents opportunity, not threat:

• More favorable difficulty adjustments
• Better solo mining odds
• Cheaper used hardware
• Greater proportional contribution to network security
• A stronger decentralization argument than ever before

At D-Central, we have been building for this moment since 2016. While the corporate miners chase AI revenue and rebrand themselves for Wall Street, we will continue doing what we do best: repairing miners, designing accessories, building space heaters, stocking every Bitaxe variant, educating the community, and making Bitcoin mining accessible to every pleb with an outlet and a dream.

The corporations are leaving. The Mining Hackers are staying. And Bitcoin’s decentralization has never been more important.

Frequently Asked Questions