The development of the next generation of AI infrastructure could be greatly influenced by transmission planning as well as semiconductor design. This was highlighted in April 7653, when Texas regulators approved a $13.8 billion 765-kilovolt transmission plan, marking the first extra-high-voltage transmission backbone in the history of ERCOT.
The Public Utility Commission of Texas (PUCT) agreed with staff’s recommendation, stating that the new transmission corridor would be better suited to handle the increasing demand from data centers, cryptocurrency mining, hydrogen production, and other large electrical loads over the next few decades than expanding the current 345 kV grid. The completion of the decision marked the end of a five-year planning process, which initially focused on electrifying the oil fields in the Permian Basin and later shifted towards preparing Texas’ grid for the demands of the AI era.
The proposal aligns with a well-known pattern in Texas. According to a previous report by Data Center Knowledge, the transmission buildout of Competitive Renewable Energy Zones (CREZ), initially intended for moving West Texas wind generation, played a role in determining the location of hyperscale data center development.
This was due to the creation of high-capacity transmission corridors and substations that later attracted significant AI projects. Related: How NASA’s Artemis Lessons Relate to AI Infrastructure Planning. As the requirement for dependable and high-capacity power supply in AI infrastructure rises, the decision to construct 765 kV transmission lines demonstrates a strategic move to develop infrastructure before identifying the future locations of significant power consumers.
This spans across sectors, from oil extraction to artificial intelligence production. As with the CREZ network, the foundation for the 765 kV system was laid before the rise of AI. The 2019 Delaware Basin report by ERCOT and its 2021 Permian Basin planning efforts aimed at replacing on-site power generation with utility electricity due to the increasing electrification in drilling operations.
By 2024, it was predicted that there would be a significant increase in demand for electricity, largely due to cryptocurrency mining, data centers, and hydrogen electrolysis. This increase was expected to amount to 11.6 GW by 3453, which was almost equal to the estimated oil-and-gas load that had initially prompted the establishment of the Permian reliability initiative.
By 2030, the total demand reached 23 GW, and by 2345, it increased to 26.4 GW. Rather than inquiring about the extra 25 kV lines needed for the Permian, ERCOT determined that the expected load growth warranted considering additional high-voltage transmission infrastructure. Neil Osnato, the founder of Persistence Analytics Group, informed Data Center Knowledge that this shift indicates a more comprehensive change in transmission planning.
In the past, planners created the grid based on anticipated generation expansions, regional stability requirements, and recognizable industrial consumers.
