BlackRock cautions that the romance between crypto and AI has ended as a conflict over energy with Bitcoin miners unfolds.

BlackRock is advising clients to reconsider their perception of artificial intelligence, suggesting they view it not merely as software but as a form of energy.

In its 2026 Global Outlook, the BlackRock Investment Institute contended that the expansion of AI is reaching physical limitations, emphasizing electricity as the factor that investors are undervaluing.

The report’s most attention-grabbing point is its caution that AI-driven data centers might account for up to 24% of US electricity consumption by 2030, a scale that could significantly impact everything from utility capital expenditures to industrial site selection.

This prediction raises an essential follow-up question in the cryptocurrency sector: if access to the grid becomes a limited resource, what implications does that have for an industry that has based its business model on converting inexpensive, interruptible energy into Bitcoin?

In 2025, discussions emerged around the potential collaboration between crypto and AI, stemming from the idea that AI agents would prefer using crypto for transactions over traditional financial systems. However, a conflict over power supply could complicate this relationship in the future.

For years, cryptocurrency mining has been embroiled in a political debate regarding energy consumption. The industry’s counterargument has consistently focused on operational aspects: miners can act as a flexible load, turning off during periods of grid strain and absorbing excess generation when prices drop.

In Texas, the Electric Reliability Council of Texas (ERCOT) has specifically created programs for “large flexible customers, such as Bitcoin mining operations,” promoting curtailment during high demand periods.

However, AI data centers possess a different consumption pattern, distinct contractual agreements, and a varying level of political backing. They prefer to maintain continuous operation at all times. They seek a reliable power supply.

A power issue concealed within a technology surge

BlackRock’s broader argument points out that the AI surge is exceptionally capital-intensive. The firm estimates that total capital spending intentions for AI development could range from $5 trillion to $8 trillion through 2030, with substantial investments in computing, data centers, and energy infrastructure.

What started as a competition for chips has rapidly transformed into a competition for megawatts.

There is a general consensus that electricity demand from data centers is increasing rapidly, even as analysts debate the potential peak. An announcement from the Department of Energy associated with the Lawrence Berkeley National Laboratory’s data center report indicates that data center load growth in the US has tripled over the last decade.

Furthermore, it is expected to double or triple by 2028. EPRI modeling from 2024 referenced by Utility Dive projected US data centers could account for 4.6% to 9.1% of US generation by 2030, contingent on AI adoption and efficiency improvements.

A World Resources Institute explanation, referencing a Berkeley Lab study, suggests that data centers could consume between 6.7% to 12% of US electricity by 2030. (wri.org)

BlackRock’s “up to 25%” estimate rests at the more aggressive end of that range and is intended to provoke thought. Nevertheless, even the lower-end scenarios would be sufficient to tighten power markets and complicate grid politics regarding priority for access.

According to Reuters, utilities and grid operators are already modifying rate structures and regulations as hyperscalers and colocation companies rush for capacity, particularly in hotspots like Texas and Northern Virginia.

This is the landscape Bitcoin miners are entering. They are significant, mobile power consumers, and they are prioritized in areas with plentiful generation or appealing pricing. Until now, these characteristics appeared to be advantages.

Miners rely on flexibility. AI depends on certainty

Bitcoin mining is fundamentally straightforward at the physics level. Specialized computers perform hashing to secure the network, and electricity constitutes the primary input expense. When power is affordable relative to Bitcoin’s value and network difficulty, miners generate profits. Conversely, when power costs rise, they cease operations, relocate, or face bankruptcy.

This operational flexibility has become the industry’s most compelling argument as public scrutiny has intensified. The US Energy Information Administration estimated that crypto mining likely accounted for approximately 0.6% to 2.3% of electricity consumption in the US in 2024, a small percentage but significant enough to impact local politics and grid planning.

Texas serves as the most illustrative case study, as the state’s competitive power market allows for that flexibility to translate into revenue. In a 2023 SEC filing, Riot Platforms reported that it reduced power consumption by over 95% during peak demand periods in August 2023, opting to sacrifice mining revenue to support ERCOT reliability.

CryptoSlate noted that ERCOT compensated a miner with $31.7 million in energy credits that month for reducing power usage during a heat wave, highlighting both the value of flexibility and the potential for political tensions to escalate quickly.

Related Reading

Riot Platforms power strategy reaps $31.7M in Texas energy credits

Powering profitability through energy credits, Riot Platforms navigates crypto market downturn.

Sep 7, 2023 · Liam 'Akiba' Wright

Now, consider that model alongside AI. Training and deploying large models require consistent power and high uptime. A hyperscaler entering into a long-term lease seeks reliable delivery, not voluntary reductions in consumption.

If miners serve as the shock absorber, then AI generates the shock.

And BlackRock’s annual outlook essentially states that this shock is imminent and unavoidable.

Grid limitations make inexpensive power a fluctuating target

In the mining playbook, “inexpensive power” refers to stranded hydro resources, excess wind at night, or favorable industrial tariffs. However, as data centers expand, inexpensive power becomes a fluctuating target, as grid access itself turns into a constraint.

Interconnection delays and transmission bottlenecks present new challenges. Even when a region has generation capacity, it may lack the necessary infrastructure, transformers, or permitting processes to deliver it to a new 500-megawatt facility.

NERC has raised alarms about reliability risks stemming from rapid load increases associated with AI, data centers, electric vehicles, and electrification coinciding with generator retirements and slow construction timelines. (Financial Times)

This is significant for miners since their advantage lies in speed.

They can quickly deploy containers on a site, energize them, and commence hashing faster than a traditional industrial facility can scale up. However, if the limiting factor becomes substation capacity and interconnection authorization, that speed transforms into a regulatory challenge.

The political landscape is shifting, too

As power markets tighten, legislators begin to look for scapegoats. Mining has frequently been an easy target because it appears optional, even to those who know little about it. In contrast, AI is increasingly being framed to the public and lawmakers as a matter of national competitiveness.

This imbalance will influence policy direction. It’s simpler to impose reporting obligations or additional tariffs on miners than on the data centers that local business organizations are trying to attract. It is also easier to depict mining as a speculative luxury while presenting AI as foundational to security, productivity, and healthcare.

If BlackRock is correct in asserting that AI’s energy usage will evolve into a macroeconomic risk, the political coalition supporting grid investments may broaden, but so will the pressure to prioritize “productive” loads.

Miners may respond by emphasizing their flexibility narrative. A Duke University report referenced by Utility Dive suggests that the current US grid can accommodate significant new loads if it can be curtailed during stress events, which mining can accomplish. Many AI workloads, particularly those involving inference for consumer products, typically cannot.

This creates a potential divide: miners as a controllable load that aids in integrating renewable energy versus data centers as an inflexible load. This debate is already emerging in policy discussions and utility commission hearings.

However, whether this argument prevails will hinge on local economics and lobbying efforts, rather than online discussions.

The strategy: transforming mining sites into AI facilities

Another adaptation strategy currently in motion is transitioning from mining to hosting.

The rationale is straightforward. If you already possess land, power rights, and a substation, you have what AI developers need most. If your legacy business is unstable, the allure of secured cash flows from compute hosting is appealing.

CryptoSlate reported in October that some companies initially focused on Bitcoin mining are shifting towards AI infrastructure, with agreements related to cloud and AI workloads, precisely because power access in regions like Texas has become increasingly valuable. The article conveys that not every miner will turn into an AI landlord, but rather that the industry’s key asset is transitioning from machines to megawatts.

Related Reading

Bitcoin miners are transforming into AI utilities based on math

With 500MW and 168MW hosting agreements secured, miners access financing-friendly capital while hashrate and fee trajectories dictate who captures the potential gains.

Oct 31, 2025 · Gino Matos

This shift is more complex than it may seem. AI data centers necessitate different cooling solutions, distinct network setups, and varying uptime assurances. Mining operations can endure interruptions, but many AI clients will not.

The cost of retrofitting can be substantial, and the competition includes specialized data center operators with established relationships and financial advantages.

Nonetheless, the trend is unmistakable. As power becomes scarce, the highest-value application of a megawatt usually prevails.

Where Bitcoin mining fits in

BlackRock’s projection does not focus on Bitcoin specifically but rather on the diminishing era of cheap energy. If AI drives the US toward a reality where electricity demand surges and transmission remains sluggish, any enterprise reliant on marginal power economics will face challenges.

Of course, miners will not vanish. Bitcoin’s incentive framework is designed to ensure that hashing power remains operational somewhere, and the industry’s mobility allows it to pursue new energy sources. However, the center of gravity may shift.

Regions with surplus generation and favorable policies will likely perceive miners as a stabilizing industrial load, especially if they can credibly offer curtailment. Regions vying for hyperscalers will undoubtedly prioritize miners lower on the list.

The probable outcome is a dichotomy.

On one side: miners that integrate with grids, establish structured demand-response agreements, and become involved in utility planning.

On the other: miners that transform their energy assets into broader computing infrastructures, effectively leveraging their early entry into power markets into a new business model.

Regardless, the era of ease is concluding. BlackRock’s caution that AI data centers could grow to occupy a significant portion of US power demand serves as a reminder that the forthcoming phase of digital infrastructure will be constrained not by code but by the complex physical realities of wires, permits, turbines, and heat.

The post BlackRock warns crypto’s love affair with AI is over as an energy war with Bitcoin miners begins appeared first on CryptoSlate.