Every large load in America is now competing for the same scarce resource: grid capacity. AI data centers, new factories, electrified buildings and households all want more power from a system that can't be expanded overnight.
Bitcoin miners used to have an easy story in that competition. They were the flexible buyer, the load that could switch off in minutes when the grid was stressed and soak up cheap surplus power when it wasn't. In some markets, especially Texas, that story earned them a warm welcome.
That welcome now comes with conditions. As CryptoSlate argued earlier this year, there is no formal deadline, but by 2027 grid operators will have enough operating data to tell which large loads behave the way they promised. Since then, Texas has written much of that test into law and market rules. For miners, the next year is a performance review.
Demand is growing faster than the grid can absorb
In its October 2026 Short-Term Energy Outlook, the U.S. Energy Information Administration projected record U.S. electricity use in both 2026 and 2027. As Carrier Management reported, total consumption is expected to climb from 4,195 billion kWh in 2025 to 4,288 billion kWh in 2026 and about 4,356 billion kWh in 2027. The EIA points to data centers dedicated to AI and cryptocurrency as major drivers.
That increase of roughly 161 billion kWh works out to about 18 GW of new round-the-clock load.
U.S. electricity consumption, billion kWh
Source: U.S. EIA Short-Term Energy Outlook, October 2026.
Commercial electricity sales, which include data centers, are forecast to pass residential sales in 2026 for the first time. Wholesale power prices are moving too: the EIA expects them to average about $52/MWh in 2026, 11% above 2025, with prices in the PJM region up about 41%.
When prices move like that, every large load gets scrutinized. Miners, whose output most people never see or use, get scrutinized first. It's the same pressure behind the AI power wall: when capacity is scarce, regulators start asking what each megawatt returns to the grid.
The miner's pitch: the load that can turn off
The miners' argument is real. Unlike an AI data center serving live customers, a Bitcoin mine has no users waiting on it. It can shut down within minutes when the grid is stressed and power back up when supply is plentiful. In principle, that makes miners exactly what grid operators have wanted for decades: a large, fast, controllable load.
Texas is where that model is most developed. According to TheEnergyMag, North American public miners held more than 5.1 GW of operational Texas capacity as of early 2026. ERCOT holds voluntary curtailment arrangements with large flexible loads, many of them miners, that cut demand when system load spikes or generation runs short.
Miners also have a strong financial reason to curtail at the right moments. ERCOT's Four Coincident Peak (4CP) mechanism sets a large customer's transmission charges for the next year based on its demand during the four highest 15-minute intervals of June through September. Analysts cited by TheEnergyMag estimate full 4CP exposure at about $50,000 per MW per year, or about $5 million a year for a 100 MW site. Avoid those four intervals and most of that cost disappears.
The catch is that nobody knows the official peak until after the fact. As the grid adds solar and huge AI loads, the hours that set the peak are getting harder to predict, and Texas regulators are reviewing whether flexible loads avoid too much cost. One option under discussion is moving from 4CP toward a 12-month method.
The catch: flexibility that depends on Bitcoin's price
A 2026 working paper on Texas mining loads, cited by CryptoSlate, found that miners do respond to wholesale prices and coincident-peak charges, but that response weakens as hashprice rises. Miners curtail most reliably when mining is less profitable, and less reliably when it's lucrative, even if the grid is under stress.
That's the core credibility problem. A grid operator can't plan reliability around a load whose flexibility shrinks whenever Bitcoin rallies.
Right now, the economics favor flexibility. According to MillionMiner, hashprice was about $40 per petahash per day in early October 2026, with Bitcoin near $86,000 and network difficulty about 15% below its 2025 record. At those margins, avoiding a peak-price hour or a 4CP interval is often worth more than the Bitcoin a site would mine. But a sharp Bitcoin rally could flip that math within weeks, which is exactly why grid operators want curtailment backed by contracts, not goodwill. I worked through that break-even math in From Hashrate to Megawatts.
Texas has written the test into the rules
In 2026, Texas moved from talking about flexible loads to regulating them. Most of these rules apply to sites of 75 MW or more, which covers the large mining campuses.
| Rule | What it requires | What it means for miners |
|---|---|---|
| Senate Bill 6: emergency curtailment | Large loads connected at transmission level after Dec. 31, 2025 must be curtailable during ERCOT firm load shed | Flexibility becomes mandatory in emergencies, not just a selling point |
| Senate Bill 6: backup generation | In an energy emergency, ERCOT can direct large loads with backup generation to deploy it or reduce load | On-site generation must be disclosed and can be called on |
| NOGRR 282 ride-through (effective Aug. 1, 2026) | New large computing loads must stay connected through defined voltage and frequency disturbances; crypto mining is named explicitly | Equipment and controls must be engineered for grid faults |
| PUCT interconnection standards (adopted Sept. 18, 2026) | Flat $100,000 study fee and financial security tied to the megawatts requested | Speculative capacity requests get expensive |
| Reliability service (in development) | A new program to pay loads that can shed on command | Verified flexibility could become a paid service |
Texas Senate Bill 6 was signed in June 2025. The Public Utility Commission adopted its interconnection standards in September 2026, and, as White & Case reported, the governor paused state permits for data center projects days later, pending ERCOT and water audits. Policymakers aren't only asking whether large loads can connect, but on what terms.
Ride-through: the reliability test miners can't skip
Controlled curtailment helps the grid. Uncontrolled disconnection does the opposite.
According to Texas Senate Bill 6 compliance analysis, ERCOT has recorded 28 events since January 2023 in which large computing loads dropped at least 100 MW during grid disturbances. When that much load disappears at once during a voltage dip, grid frequency and voltage can swing sharply, exactly what operators are trying to prevent.
ERCOT's NOGRR 282 rule responds directly. New large computing loads must ride through defined disturbances. If a site fails, it has 90 days to investigate, 90 days to submit a fix and 180 days to complete it, and ERCOT can order disconnection if the risk is immediate. ERCOT is also drafting limits on how fast large loads can swing their power use.
Compliance isn't cheap. Industry estimates run from $500,000 to $1 million per MW, and more for some battery-based fixes. But battery storage can do double duty, supporting ride-through while also smoothing ramps and earning revenue when the grid is calm.
The AI pivot is thinning the herd
The miners making this case are also changing who they are. According to MillionMiner, public miners moved about 75 EH/s of capacity away from Bitcoin in the first half of 2026. IREN plans to exit Bitcoin mining by the end of 2026, Cipher by the end of 2027, and Hut 8 doubled its Beacon Point campus in Texas to 704 MW under AI contracts.
Yet Bitcoin's network hashrate remains near 950 EH/s, close to its record, as other miners fill the gap. CryptoSlate estimated the U.S. held about 37.5% of global hashrate in early 2026.
The result: the miners still mining at scale in 2027 will be the ones that chose Bitcoin, not the ones that simply haven't left yet. Their case for grid access will rest almost entirely on flexibility. I covered the business side of that choice in From Hashrate to Megawatts.
Miners vs. AI data centers vs. everyone else
Every large load now has to justify its claim on the grid:
- AI data centers: firm, high-uptime demand framed as strategic infrastructure. Hard to curtail, but politically favored.
- Bitcoin miners: interruptible and price-sensitive, but valuable only if that flexibility is dependable.
- Manufacturers: inflexible, politically protected and increasingly vocal about rising capacity charges.
- Households: the center of ratepayer politics, as North Carolina and other states debate who pays for new generation.
What miners need to prove by 2027
- 1.Verified curtailment: metered, contracted demand response that happens on schedule, regardless of hashprice.
- 2.Ride-through performance: sites that stay connected through faults instead of tripping offline in bulk.
- 3.Renewable alignment: data showing mining load absorbs surplus wind and solar instead of competing for power in tight hours.
- 4.Transparency: accurate forecasts, disclosed backup generation and clean interconnection data, now required in Texas.
Two paths forward
The bear case: miners fall short on ride-through or curtailment, interconnection reviews tighten, and pure-play mining sites lose value next to operators with AI leases. Mining becomes the first load cut on a crowded grid.
The bull case: miners turn flexibility into a documented, paid grid service through programs like Texas's coming reliability service. Flexible megawatts earn a premium based on grid behavior, not Bitcoin's price.
What this means for developers and operators
If you're developing or operating a large flexible load, count the cost of proving it. Budget for ride-through engineering, metering and curtailment controls. Model your 4CP exposure and how it changes if Texas moves to a 12-month method. And compare running versus curtailing at different hashprice levels before you sign a power contract. This is part of the broader question I explored in Who Should Pay for the AI Power Boom.
The bottom line
The AI power wall has made every megawatt political. Utilities and regulators now ask every large load the same question: what does the grid get in return? Miners have a strong answer, flexibility, but in 2026 Texas started requiring proof. The miners that can show it, on the hottest afternoons and when Bitcoin is rallying, will have a better claim to grid access than almost any other large load.
Chris Kalowes
Founder, WattThe?!
Energy intelligence. Simplified.
