ERCOT, the Texas grid operator, drew more than 700 participants to a workshop this week where it rolled out a draft plan to overhaul how large loads connect to the state's grid. The backdrop is staggering: more than 250 gigawatts of large-load requests now sit in the interconnection queue, roughly three times the operator's current peak demand of about 85 GW, with the bulk coming from AI-focused hyperscale data centers rather than crypto miners. The operator described the existing process as structurally unsuited to a queue of that scale, and stakeholders broadly agreed.
The current process — bilateral, one-at-a-time transmission impact studies — has buckled under the volume. ERCOT staff have publicly acknowledged that some applications submitted in 2024 are unlikely to receive even preliminary studies before 2027 under the legacy framework, and a meaningful fraction of the queue is widely understood to consist of speculative applications that may never materialize. The new draft replaces sequential studies with system-wide batch reviews on a six-month cadence, allowing the operator to model new loads against the full set of pending applications rather than treating each one as an isolated change to the network.
Alongside the long-term redesign, ERCOT unveiled a transitional "Batch Zero" phase to triage the most advanced projects already in the existing queue. Roughly 8.2 gigawatts of pending applications may be subject to reassessment, with priority going to projects that have already secured land, signed firm power contracts, and obtained the air, water, and zoning permits needed to actually break ground. Speculative applications — what staff have informally called "queue squatters" — face a meaningfully higher bar, and some are expected to be culled outright. The operator framed the cull as essential to giving real projects realistic in-service timelines.
For Bitcoin miners, the most consequential signal was ERCOT's open interest in formally recognizing flexible loads with controllable demand profiles as a separate category. Miners can ramp from full draw to zero in under fifteen minutes; AI training clusters cannot, because checkpointed training runs that get interrupted lose hours or days of progress on multi-million-dollar GPU time. ERCOT staff floated the idea of a tiered interconnection process where flexible-load applicants would receive faster studies and reduced transmission upgrade cost responsibility, on the explicit condition that they accept binding curtailment commitments enforced through automated dispatch infrastructure rather than self-reported compliance.
Industry response was cautiously positive. The Texas Blockchain Council and several public miners filed comments supporting the flexibility tier, while the Data Center Coalition — which represents the AI hyperscaler interests — pushed back on any framework that would systematically advantage flexible loads over inflexible ones. The political dynamics in Austin tilt slightly in the miners' direction: ERCOT's reliability mandate is hard to argue against, Bitcoin mining's documented role as dispatchable load during the past three winter storms has given the lobby a credible operational record, and the 1.4-gigawatt coordinated curtailment during last summer's stress event remains a frequently cited reference point in regulatory testimony.
The implications for grid economics are large. Texas already sources roughly 38% of global Bitcoin hashrate and is the leading U.S. AI data center market by some distance. How ERCOT's interconnection rules treat the two industries over the next eighteen months will substantially influence both sectors' long-term geographic distribution and the multi-billion-dollar capital allocation decisions sitting behind the queue. The next milestone is a formal stakeholder-comment window followed by board action; the substation, transmission, and grid blueprint that emerges will be read carefully by every operator with capacity ambitions in the state. Final rules are expected in the second half of the year.