Can Bitcoin Miners and AI Data Centers Share the Same Power Grid? BlackRock Says the Love Affair Between Crypto and Energy Is Cracking

The honeymoon is over. What began as a narrative-driven love affair between cryptocurrency and artificial intelligence in 2025—built on the theory that AI agents would fuel demand for blockchain payments—now faces a hard constraint: power. BlackRock’s 2026 Global Outlook dropped a bombshell for the industry: AI-driven data centers could consume as much as 24% of US electricity by 2030. That’s not just a headline. It’s a pivot point that reorders who gets to plug in and who gets shut out.

When Baseload Becomes a Weapon

Here’s the fundamental tension: Bitcoin mining was built for flexibility. Crypto miners are the grid’s shock absorber—they power down when demand spikes, they switch on when renewables are cheap and plentiful. In Texas, ERCOT explicitly designed programs for “large flexible customers, such as Bitcoin mining facilities” precisely because miners can curtail without collapsing their operations.

But AI data centers come with a completely different profile. They demand baseload power. They don’t want to turn off, ever. They want constant, uninterrupted supply. Think of it this way: miners are willing players in grid management. AI is an unmovable load that only consumes more.

The contrast becomes sharper when you look at actual numbers. Riot Platforms told the SEC it curtailed power usage by over 95% during peak demand in August 2023, and ERCOT paid them $31.7 million in energy credits that month for powering down during a heat wave. Try asking a hyperscaler running training models to do the same. It won’t happen.

The Math Behind the Power Crunch

BlackRock isn’t making this up. The firm estimates $5 trillion to $8 trillion in total capital spending for AI infrastructure buildout through 2030, with massive outlays on compute, data centers, and transmission. What started as a chip race became a megawatt race overnight.

Data center electricity demand has already tripled over the past decade, according to Department of Energy analysis tied to Lawrence Berkeley National Laboratory research. Projections show it doubling or tripling again by 2028. The spread of estimates is wide—EPRI modeling suggests data centers will hit 4.6% to 9.1% of US generation by 2030, while World Resources Institute cites 6.7% to 12% depending on AI adoption and efficiency. BlackRock’s aggressive “up to 25%” framing sits at the outer edge, but even the conservative scenarios mean grid stress.

And here’s what matters for miners: grid stress kills optionality. When power becomes the bottleneck, speed becomes irrelevant. Bitcoin mining’s historical advantage—containerizing operations and spinning up hash power faster than traditional industry—evaporates if the constraint shifts from generation to transmission infrastructure. Interconnection queues and substation capacity become the gating factor, not fuel availability.

The Political Economy Gets Ugly

Lawmakers don’t wake up worrying about Bitcoin mining. They wake up worrying about whether American AI companies can compete with China. That asymmetry in political salience is everything.

When grid capacity tightens, miners become easy targets. Mining feels optional. It feels speculative. In contrast, AI gets framed as essential to national competitiveness, defense, productivity, and medicine. It’s easier for a regulator to impose additional tariffs or reporting requirements on miners than on the data center the local chamber of commerce just recruited with tax incentives.

This is already playing out. Reuters reported that utilities and grid operators are adjusting rate structures and rules as hyperscalers scramble for capacity in hotspots like Texas and Northern Virginia. Miners competing for the same real estate face an uneven playing field.

The Flexibility Argument: Mining’s Last Defense

Miners are trying to fight back with a sophisticated counter-narrative: position mining as the controllable load that actually helps integrate renewables and stabilize grids. A Duke University report cited by Utility Dive argues the existing US grid can handle significant new load growth if that load can be curtailed during stress events. Mining can. Most AI inference workloads serving consumer products cannot.

It’s a wedge argument designed to appeal to utility commissions and state-level energy planners. But whether it wins depends on local economics and actual lobbying power, not internet arguments.

The Hedge: Become the Infrastructure You’re Competing Against

Some miners are already pivoting. CryptoSlate reported in October that firms originally focused on Bitcoin mining are shifting toward AI infrastructure hosting, with deals tied to compute and cloud workloads. The logic is ruthless: if you own land, long-term power rights, and substation access in a place like Texas, you have what AI developers need most. Contracted cash flows from compute hosting beat the volatility of mining revenue.

This isn’t a universal move. Retrofitting mining operations for AI hosting is harder than it sounds. AI data centers require different cooling, different network topology, different uptime guarantees that mining operations weren’t engineered for. The competition includes specialist data center operators with deeper relationships and better financing. But the trajectory is clear: when power becomes scarce, the highest-value use of a megawatt tends to win.

The Barbell Future

Bitcoin mining won’t disappear. The network’s incentive structure ensures hash power comes online somewhere. Mining’s global mobility means it can chase new energy pockets. But the center of gravity in the industry will shift.

One path: miners integrate with grids and sign formal demand-response agreements, positioning themselves as stabilizing industrial load. They lean into the flexibility narrative and become utilities’ partners in grid management.

The other path: miners leverage their early arrival in power markets to pivot toward broader compute infrastructure. They arbitrage their energy positions into a new business model.

Either way, the era of cheap abundance is over. The next phase of digital infrastructure—whether Bitcoin or AI—won’t be constrained by code or computing power. It will be constrained by wires, permits, turbines, and the unglamorous physics of how electricity actually gets delivered. That’s the world BlackRock is warning about, and it’s reshaping who wins and who loses in the infrastructure sweep.