Bernstein: Bitcoin miners poised as key AI infra suppliers

Bitcoin miners are increasingly positioning themselves as pivotal players in the AI infrastructure supply chain, leveraging their control of sizable power capacity and data-center real estate to support surging demand for AI workloads. A fresh Bernstein analysis shows publicly traded miners collectively plan more than 27 gigawatts of power capacity and have disclosed AI-related agreements totaling over $90 billion, covering about 3.7 gigawatts with hyperscalers, neocloud providers and chipmakers. The finding adds a new dimension to the industry’s post-halving trajectory, suggesting energy and site access could become the true bottlenecks in scaling AI computing, even as crypto mining undergoes a notable pivot toward AI-focused data centers and high-performance computing facilities.

Meanwhile, a RAND research brief released last week estimates the United States could add roughly 82 gigawatts of net available capacity by 2030, underscoring a broader backdrop of expanding demand for data-center-grade power. Bernstein emphasizes that the real constraint now is electricity access—grid interconnections and approvals can take years, complicating plans to scale AI infrastructure at pace. In practice, the wait times for securing a gigawatt of power can stretch to about 50 months across states, with even growth-friendly jurisdictions such as Texas applying batch-review processes to manage interconnection queues and resource loads. The combination of regulatory scrutiny and local opposition to large-scale data centers further compounds these delays, in Bernstein’s view giving miners an edge due to their existing, grid-connected sites and experience running high-density computing facilities.

With AI demand rising, the report frames Bitcoin miners as potential accelerants for AI infrastructure rather than mere participants in a crypto cycle. The authors note that the bottleneck has shifted from silicon to electricity, a change that could reshape strategies across the crypto and broader tech infrastructure sectors.

The analysis arrives amid a broader narrative that the so-called AI supercycle is not only about chip technology or cloud-scale compute, but also about who can reliably provide the energy and real estate required to run demanding AI workloads at scale. The piece links to prior coverage on how miners are moving beyond traditional Bitcoin production to build data-center ecosystems capable of hosting AI-related infrastructure and computing workloads.

Key takeaways

• Miners control a planned power portfolio exceeding 27 GW and have disclosed more than $90 billion in AI-related agreements covering about 3.7 GW with hyperscalers, neocloud providers and chipmakers, according to Bernstein.
• Access to electricity has become the primary scaling constraint for AI data centers, with grid interconnection queues and permitting delays stretching into multi-year timelines in several states.
• RAND projects a significant growth path for US capacity, estimating around 82 GW of net available capacity could be added by 2030, highlighting a larger macro backdrop for AI infrastructure expansion.
• The regulatory environment and local opposition to large data centers are contributing to delays, reinforcing the advantage for miners already operating grid-connected facilities.
• Miner economics are evolving: after the 2024 halving reduced mining rewards, several players are expanding into AI data centers and high-performance computing, with Soluna Holdings reporting a substantial rise in data-center hosting earnings, while IREN is cited as a prime pivot candidate thanks to Microsoft-backed AI agreements.

AI infrastructure takes the lead, while electricity remains the hurdle

Bernstein’s analysis paints a picture of miner-turned-AI infrastructure players extending beyond their core Bitcoin mining activities. After the 2024 halving compressed mining margins, the sector has increasingly pursued revenue diversification through AI data centers and high-performance computing facilities. The emphasis is shifting from raw hashing power to the ability to secure reliable power and proximity to robust data-center ecosystems—assets that miners already command through long-standing grid connections and experience managing complex, dense computing environments.

The practical implication for investors and builders is clear: the value proposition for miners hinges less on the price of Bitcoin and more on their capacity to unlock and monetize AI-ready energy and real estate. The interconnection bottleneck is no longer a theoretical risk but a real choke point that can slow or derail expansion plans. In this context, utility providers’ approval processes, capacity queues and the pace of grid upgrades become material factors shaping the pace of AI infrastructure deployment. This dynamic helps explain why miners with established infrastructure networks may enjoy a structural advantage as AI workloads proliferate across industries.

Real-world pivots: from mining to AI clouds and data centers

The Bernstein study spotlights concrete examples of diversification beyond traditional crypto mining. Soluna Holdings, for instance, reported a meaningful uptick in first-quarter revenue, driven largely by its data-center hosting business rather than crypto mining. The shift mirrors a broader pattern among miners seeking recurring, sizable revenue streams tied to AI-ready facilities rather than volatile mining rewards alone.

Another prominent example cited by Bernstein is IREN, which is viewed as well-positioned to pivot toward AI infrastructure following multibillion-dollar agreements with Microsoft. The premise is simple: if miners can leverage existing sites and operational expertise to house AI compute and related services, they may unlock new growth channels that complement, or even supplant, traditional mining economics over time.

These moves are not merely opportunistic. They reflect a strategic recalibration in response to both market pressures and regulatory realities. By leveraging grid-connected sites and building AI-capable data centers, miners could become essential partners in AI value chains—providing power, cooling, and space for AI cloud services, while also contributing to the resilience and redundancy of AI compute ecosystems.

For investors, the takeaway is that AI infrastructure demand is not a standalone trend but a potential economic expansion path for miners with the scale and site access to support large, power-intensive deployments. It also underscores a broader market shift: the traditional crypto cycle may increasingly ride on AI-driven demand for compute and data-center capacity, rather than price dynamics alone.

What remains uncertain, however, are the policy and regulatory trajectories across different geographies and how quickly grid operators can modernize the interconnection process. The RAND projection of 82 GW of additional capacity by 2030 provides a bullish backdrop, but the pace at which administrators authorize new connections will be crucial. The coming years could determine whether the mining-to-AI infrastructure pivot achieves its intended scale or encounters persistent friction in the form of permitting delays and local opposition.

Beyond the headline figures, the evolving economic model invites a closer look at how specific players balance energy costs, capital expenditure for data-center facilities, and revenue from AI-related services. The Soluna and IREN cases illustrate how diversified revenue streams—from hosting to cloud-style AI offerings—may become a backbone for miner profitability, particularly as traditional block rewards continue to adjust post-halving cycles.

Additionally, the broader AI hardware supply chain remains a critical factor. As miners court partnerships with hyperscalers, cloud providers and chipmakers, the question becomes not only who can secure the most power but who can integrate seamlessly with AI platforms and meet reliability standards essential for enterprise-grade compute workloads. In this sense, the Bernstein analysis casts miners as potential accelerants for AI infrastructure growth, provided they can navigate energy and regulatory complexities with the same efficiency they apply to data-center management.

In short, the convergence of Bitcoin mining and AI infrastructure signals a meaningful shift in how digital asset infrastructure assets are valued. It points to a future where energy access, site strategy and long-term power commitments may determine which players lead in AI-enabled compute—and which ones struggle to scale in the face of interconnection bottlenecks and policy headwinds.

Readers should watch how grid operators, regulators and utility providers respond to this evolving landscape, as well as how mining firms optimize their asset portfolios to capitalize on growing AI demand while managing the risk profile that comes with long interconnection timelines and the complex economics of data-center deployments.

This article was originally published as Bernstein: Bitcoin miners poised as key AI infra suppliers on Crypto Breaking News – your trusted source for crypto news, Bitcoin news, and blockchain updates.