
Field Notes
The Geostrategic Grid: How AI Data Center Power Demand Is Rationing National Electricity Capacity
On June 18, 2026, the Federal Energy Regulatory Commission ordered all six major US grid operators, PJM, MISO, SPP, CAISO, ISO-NE and NYISO, to justify their existing large-load interconnection rules within 60 days or propose reforms.[1] The same week, the Public Utility Commission of Texas approved a new batch-processing framework for a queue that had grown to 438 gigawatts of requested large-load capacity, close to 90 percent of it data centers.[2] Neither action targeted a single hyperscaler site. Both were forced by the aggregate weight of demand that grid planners did not model five years ago, and both turned a utility engineering problem into a regulatory and geopolitical risk one in the same news cycle.

The Geostrategic Arithmetic Behind the Interconnection Queue
Grid operators are not reacting to a marginal demand blip. The IEA's Electricity 2026 report puts global electricity demand growth at 3.6 percent annually through 2030, about 50 percent faster than the prior decade's average, and attributes roughly half of the incremental demand growth in the United States to data center expansion.[1] That arithmetic is now showing up as formal intervention rather than internal utility memos.
- FERC's show-cause orders. The Commission gave the six regional grid operators 60 days to justify or reform tariff rules on co-location, behind-the-meter generation and expedited transmission service, plus 30 days to report on generation adequacy. FERC Chairman Laura V. Swett framed the goal as transforming "the way large energy users access the grid" while protecting consumers from cost-shifting.[1]
- Texas's 438-gigawatt queue. ERCOT's large-load interconnection requests have grown so fast that the state built an entirely new batch-review process, with a 75-megawatt threshold for study and the first cohort not expected to reach a final transmission plan before fall 2027.[2]
- PJM's load forecast. The country's largest grid operator projects 55 gigawatts of new large-load growth by 2030 and 100 gigawatts by 2037, with utilities in its footprint now committed to serve roughly twice the new load their planned generation can support.[3]
- Ireland's reversal. After a moratorium in place since 2021, the Commission for Regulation of Utilities published a final connection policy on December 12, 2025, reopening the grid to new data centers on condition they install on-site generation or battery capacity, source at least 80 percent of demand from new renewables, and export power back to the grid on request. Data centers already consume close to 25 percent of Ireland's electricity, up from about 5 percent a decade ago.[4]

What Ireland, Malaysia and Texas Reveal About Geostrategic Power Allocation
None of these decisions were made by a technology company choosing a site. They were made by regulators and grid operators allocating a scarce national resource among competing claimants, which is precisely the definition of a geopolitical risk decision rather than a commercial one.
- Grid capacity is now a sovereign allocation problem. Ireland's policy does not ask whether a data center should connect. It asks what the data center must contribute back to the system to earn the connection, effectively pricing grid access as a strategic good rather than a utility service.[4]
- Cross-border power politics is already active in Southeast Asia. Malaysia's data center buildout in Johor is projected to add 59 terawatt-hours of demand by 2030, roughly the Philippines' entire annual electricity consumption, concentrated in the Johor-Singapore Special Economic Zone. Malaysia's government has paused approval of new non-AI data centers over water and power supply constraints even as it courts AI-linked investment, and Chinese state-linked firms including PowerChina and Huawei are now embedded in the grid buildout meant to support that demand.[5]
- The condition-setting playbook is becoming standard, not exceptional. Ireland's onsite-generation and renewable-sourcing conditions, and FERC's cost-shifting and co-location safeguards, are converging on the same template: access to scarce capacity in exchange for the load itself absorbing part of the reliability and cost risk it creates.[1][4]

What Geostrategic Grid Risk Management Requires Now
Energy-sector risk and security leaders have generally filed data center demand under capacity planning. The evidence from 2025 and 2026 argues it belongs in the same geopolitical risk register as chokepoint disruption and sanctions exposure.
- Rerank infrastructure disruption as a near-term risk, not a background one. The World Economic Forum's January 2026 analysis with Zurich Insurance Group found that critical infrastructure disruption ranks only 23rd among risks over the next decade, despite ranking fifth as a direct conflict target over the next two years, a gap the report calls a structural oversight in how organizations weigh the risk.[6]
- Treat interconnection queue position as an intelligence input. A 438-gigawatt queue or a 100-gigawatt forecast is a forward-looking indicator of where transmission investment, political attention and potential rationing will concentrate. Risk and operations leaders in energy-adjacent sectors should track queue movement the way they track sanctions lists.
- Watch the attack surface expand with the grid. Cyberattacks on US utilities rose from 689 to 1,162 incidents in 2024, a 70 percent increase, and roughly 60 percent of attacks on critical infrastructure are now attributed to nation-state actors.[7] Every new large-load interconnection point, substation and transmission upgrade tied to data center buildout adds to that surface.
- Track cross-border power arrangements as a siting variable, not an engineering footnote. Johor's dependence on Singapore-adjacent power infrastructure and China's parallel investment in Malaysian grid capacity show that data center siting decisions increasingly carry the same cross-border dependency risk as semiconductor supply chains.[5]
Fortius Intel note: The companies exposed to this shift are not only the hyperscalers choosing where to build. Utilities, grid operators and the industrial and financial customers who share their transmission capacity are now competing for the same scarce interconnection slots, and the allocation rules being written in 2026 will define who gets power first for the next decade.
Methodology: Analysis draws on the IEA's Electricity 2026 report (2026), FERC's June 18, 2026 order to six US regional transmission organizations, Public Utility Commission of Texas and ERCOT filings on the large-load interconnection framework (June 2026), the Commission for Regulation of Utilities' Large Energy User Connection Policy (December 12, 2025), reporting on Malaysia's Johor data center buildout, the World Economic Forum's January 2026 critical infrastructure risk analysis with Zurich Insurance Group, and TTMS's energy-sector cybersecurity threat data. All cited sources are publicly available.
Footnotes
1 Federal Energy Regulatory Commission, "FERC Launches Aggressive Targeted Action to Speed Large Load Integration," FERC, June 18, 2026. Show-cause orders issued to PJM, MISO, SPP, CAISO, ISO-NE and NYISO with a 60-day deadline to justify or reform interconnection tariffs. Available at https://www.ferc.gov/news-events/news/ferc-launches-aggressive-targeted-action-speed-large-load-integration. See also IEA, Electricity 2026, International Energy Agency, 2026, on global electricity demand growth of 3.6 percent annually through 2030 and data centers accounting for roughly half of US demand growth. Available at https://www.iea.org/reports/electricity-2026/executive-summary and https://www.iea.org/reports/electricity-2026/demand.
2 Utility Dive, "Texas, facing 438 GW queue, approves initial large-load interconnection process," Utility Dive, June 2026. Public Utility Commission of Texas approval of a batch-review framework, 75 MW study threshold. Available at https://www.utilitydive.com/news/texas-facing-438-gw-queue-approves-initial-large-load-interconnection-pro/823367/.
3 White & Case LLP, "PJM proposes to carve out new services for co-located data centers," White & Case, 2026. PJM forecast of 55 GW of new large-load growth by 2030 and 100 GW by 2037. Available at https://www.whitecase.com/insight-alert/pjm-proposes-carve-out-new-services-co-located-data-centers.
4 EnergyConnects, "Ireland Ends Moratorium on New Power Links to Data Centers," EnergyConnects, December 2025. Commission for Regulation of Utilities' final Large Energy User Connection Policy, published December 12, 2025, requiring on-site generation, 80 percent renewable sourcing, and grid export capability. Available at https://www.energyconnects.com/news/utilities/2025/december/ireland-ends-moratorium-on-new-power-links-to-data-centers/.
5 China Global South Project, "China Steps In as Malaysia's Data Center Surge Puts the Power Grid to the Test," China Global South Project, 2026. Malaysia data center demand growth of 59 TWh by 2030, Johor development pause, and Chinese firms' role in grid buildout. Available at https://chinaglobalsouth.com/analysis/china-malaysia-data-centers-power-grid/.
6 World Economic Forum, "Why critical infrastructure risk deserves more attention in 2026," World Economic Forum, in partnership with Zurich Insurance Group, January 14, 2026. Infrastructure disruption ranked 23rd among 10-year risks despite ranking fifth as a two-year conflict target. Available at https://www.weforum.org/stories/2026/01/critical-infrastructure-global-risks-2026-zurich-insurance/.
7 TTMS, "Guide to Cybersecurity Threats in the Energy Sector," TTMS. US utility cyberattacks rose from 689 to 1,162 incidents in 2024, a 70 percent increase; approximately 60 percent of critical infrastructure attacks attributed to nation-state actors. Available at https://ttms.com/guide-to-cybersecurity-threats-in-the-energy-sector/.
About the author
Jay Bimbrah, Co-Founder & COO. A former Scotland Yard counter-terrorism investigator, Jay has advised EMEA tier-1 banks and Lloyd's market firms on distinguishing real exposure from theoretical risk.