The OMS Transmission Planning and Resources Work Groups (OMS Work Groups) provide this feedback to MISO regarding the Long-Term Load Forecast (LTLF) Workshop. This feedback is from OMS work groups and does not represent a position of the OMS Board of Directors.

What improvements would make the forecast outputs more actionable or transparent?

• The OMS Work Groups support greater transparency regarding the data sources, assumptions, and methodologies used in the LTLF. It would be helpful for workshop materials and resulting white paper to include an appendix documenting key methods, data sources, references to stakeholder surveys, and other inputs used in developing the forecast.
  • The white paper generally references broad source categories, but provides limited citations to specific publications, datasets, and vintages. The workshop slides appear to contain more detailed references (e.g., specific references to FERC Form 715 DER data). MISO should strive to include more specific source citations in the white paper where practical, particularly when data sources materially influence forecast outcomes.
• The OMS Work Groups encourage MISO to maintain consistency in data sources and forecasting methodologies over time. When changes are made, stakeholders should be informed of the changes, the rationale for the changes, and the expected impact on forecast results.
• MISO should provide more clarity on the stakeholder submitted data (the black-dotted line):
  • Was this line updated with the final forecast submissions collected in 2026 in advance of the PRA? The workshop slides indicate it is “preliminary,” and the white paper does not specify.
  • Why is the black-dotted line in 2026 smooth, whereas the corresponding line in the December 2024 Long-Term Load Forecast show more variation by year?
• Can MISO provide more detail about the Kevela analysis, and how their results changed the LTLF?
• The OMS Work Groups express caution when developing data center and BTMG assumptions. From the workshop, it sounded like MISO was contemplating the possibility that reported data center load could be over- or under-estimated due to apparent data center interest in BTMG.
  • If so, we observe that initial experiences with data centers indicate interest in BTMG varies widely. Unless forecasts are provided that explicitly state large loads would incorporate BTMG, we would not suggest making assumptions about BTMG interest.

Which areas of the load forecast should MISO prioritize for enhancement?

N/A

Are there emerging trends or technologies in your footprint that MISO should explicitly incorporate?

• The OMS Work Groups express interest in seeing a low-end estimate for large load that contemplates more substantial limiting factors to data center growth such as supply chain issues, workforce issues, and/or regulatory lag.
  • Another aspect to consider would be the maturation of the industry itself.
• If data center development does not follow the normal S-curve development as an industry and finds ways to be more efficient with input/outputs, we can expect the industry to mature like anything else and continued growth should not be overloaded. MISO seems to be factoring in this consideration, but if the initial growth is overstated, the forecasts may still be too high.

OMS Work Group members questioned whether data centers will continue growing at 17.9% to 18.2% through 2046. Figure 15 indicates growth assumptions of 91% CAGR for 2026–2030, 6% for 2030–2035, and 1.8% thereafter. Similar uncertainty applies to other rapidly evolving categories, such as EVs.

What additional collaboration opportunities could strengthen alignment between MISO and stakeholders?

OMS Work Group members continue to explore potential and intended uses of the LTLF, which may impact the need for accurate inputs or various sources of data (e.g., utility-only, third-party, disclosure requirements) may be needed depending on use output.

What aspects of MISO’s uncertainty approach need greater transparency, including any areas where forecasts may consistently over- or under-estimate expected conditions?

The OMS Work Groups request further detail on why adjustments need to be made for particular factors or trends. This recommendation recognizes that the impact of some of these factors may already be appearing in historical data in which case adding further medications might be unnecessary.

• We request more information about the role of Landgate including the specific tasks they completed and how they were selected.
• Did MISO consider any potential conflict-of-interest issues with having a company that sites- and finds capacity for-large load involved in this work and report?

Are there any timing considerations we should be aware of as MISO establishes a predictable cadence for releasing load forecasts?

• Having predicable out-year forecasts for class-level accreditation values is critical. As those values become outputs of MISO’s Futures study and the long-term load forecast is an input into the Futures study, we recognize a need for a predictable cadence to accomplish that objective.
  • While every LTLF does not necessarily inform each Futures initiative, we see a need for MISO to discuss needs and objectives with stakeholders about expectations around the release of out-year class-level accreditation values.

Finally, the OMS Work Groups note that it would increase clarity for stakeholders if naming conventions for these studies could be more consistent. For example, it appears the 2026 LTLF should be named the 2025 LTLF if following past patterns for how MISO names these studies.

The Environmental Sector supports MISO’s Long-Term Load Forecast effort, including the regularity of updates (approximately annually) given the volatile and unprecedented load growth landscape.

Q: What improvements would make the forecast outputs more actionable or transparent?

A: The report should include an explanation of how this forecast informs MISO’s internal priorities

The Executive Summary workshop slide states, "As part of its Reliability Imperative, MISO is enhancing long-term load forecasting to inform long-lead decisions." Which of MISO's long-lead decisions will this forecast inform? Supporting the next iteration of the Futures exercise and improving the forecast for next year cannot be the only downstream impact of this consequential forecast.

The Sector strongly urges MISO to:

• Ensure the timelines remain on track for the next tranche of long-range transmission to address insufficient grid hosting capacity,
• Significantly increase the Generation Interconnection staff - with competitive salaries and manageable workloads - to address the interconnection bottleneck, including the Surplus Interconnection path which is also becoming overwhelmed by the volume of new applicants, and
• Provide resource planners certainty on the accredited value of Storage, External Resources, and interregional HVDC Facilities to enable needed investment in flexible and interregional attributes.

A: The report should provide more granular information to improve planning value to stakeholders

While MISO currently withholds LRZ forecasts to protect data center confidentiality, this lack of granular data limits planning value. Rather than treating this as a permanent restriction, we urge MISO to find a middle ground. Potential solutions include aggregating data to a minimum threshold to mask individual projects or sharing the data under existing non-disclosure (CEII) frameworks. These steps would provide essential local insights while still protecting sensitive project information. 

MISO should provide underlying data in appropriately aggregated and reviewable format (e.g., excel). While the white paper explains how MISO forecasts, the lack of granular data (such as projections by zone (LRZ), weighting factors for the various assumptions, and the use of third-party sources) prevents stakeholders from providing more meaningful feedback. Given that data center projections have doubled since 2024 while hydrogen and EV goals have shifted, stakeholders should be able to evaluate the quantitative assumptions that led to these outcomes.

Q: Which areas of the load forecast should MISO prioritize for enhancement?

A: Scenarios structured on distinct drivers

While we appreciate the increased transparency in the 2026 LTLF, the three trajectories remain too focused on a single variable of data center realization rates. This creates a narrow range of outcomes that fails to capture the true uncertainty of the next 20 years. We would recommend structuring scenarios around distinct drivers such as: Unconstrained Data Center Growth; Efficiency & Moderation (e.g., technological gains that lower power demand over time); or Market Contraction (e.g., delayed realization due to monetization failure, economic downturn, or policy reversal). By using more stable drivers (like residential demand) as secondary sensitivities, MISO can evaluate specific scenarios rather than bookends.

Additionally, the white paper notes that “historical load comparisons in Figure 10 show that, despite member forecasts over the past 5 years and 2024 LTLF, signaling varying magnitude of growth, the actual demand has remained flat.” Accurately predicting future load growth is complex — prior experience with emerging technology forecasts suggests adoption can follow an S-curve that initial projections may miss. MISO should consider efficiency gains in AI hardware as an additional moderating factor.

A: Behind the meter generation

MISO indicated during the April 13 workshop that 30-60% of data center load may be backed by on-site generation. If these resources are not registered in MISO markets or net load metering, they could lead to systematic overestimation of coincident peak demand.

A: Extreme weather scenarios

We appreciate the partnership with Kevala and encourage MISO to prioritize reconciling the estimates through Kevala with the 1.6 GW 2021 AEG forecast and the 11.3 GW 2025 OMS survey. We request MISO to prioritize extreme weather scenarios to understand how operations translate into the long-term forecast. For example, with increasing electrification of heating and the structural shift toward higher system load factors driven by data centers, winter peak dynamics are becoming more consequential, and should receive dedicated treatment in future forecasts.

Q: Are there emerging trends or technologies in your footprint that MISO should explicitly incorporate?

A: Anticipating the specific profiles for varying AI loads will be important to understand coincident peak demand[1]

Frontier AI training demand is highly concentrated among a small number of labs. Projections suggest individual training runs could reach 4–16 GW by 2030, but the number of facilities required to support frontier training at that scale remains small in absolute terms relative to the broader hyperscale buildout driving MISO's pipeline. MISO should assess how much of its data center pipeline reflects this concentrated frontier training demand versus the much larger and more distributed inference and cloud workload market. MISO should consider whether the "1+ GW campus" is a specialized 2027–2028 phenomenon or a universal standard for the forecast period.

Additionally, the ~93-94% load factor assumption may overstate energy intensity by conflating two distinct workloads. A growing share of AI activity consists of inference workloads that are bursty and request-driven, rather than the constant high-density load profile of frontier training facilities. These workloads do not require gigawatt-scale campuses and are increasingly served by smaller, distributed facilities co-located within existing cloud infrastructure. MISO should consider differentiated load factor assumptions that distinguish frontier training hubs from distributed inference deployment, which may exhibit materially different coincident peak demand profiles.

A: MISO should assess supply chain constraints for AI data center build out[2]

The AI data center build out has a very concentrated supply chain. Multiple chip designers (dominated by Google and Nvidia) create the chip blueprints, but the manufacturing is very difficult and the majority is performed by TSMC (logic dies and CoWoS) and SK Hynix, Samsung, and Micron manufacture the HBM (high-bandwidth memory) stacks. This concentration means that any bottlenecks can propagate quickly. We recommend MISO assess whether supply chain constraints on critical components could limit the pace of data center buildout assumed in the Current and High trajectories.

A: MISO should consider the forecasted costs of fossil gas, delivered fuels, gasoline, and electricity to anticipate fuel-switching tipping points

The Sector recommends MISO consider the forecasted costs of fossil gas, delivered fuels, gasoline, and electricity to anticipate fuel-switching tipping points. For example, the rising costs of electricity will shorten the ROI timeline for rooftop solar or other efficiency investments, potentially accelerating adoption. The rising gas prices will accelerate the gas-to-electric transition for residential and commercial heating. There is already evidence that rising gasoline prices driven by geopolitical instability may accelerate electric vehicle adoption. To the extent that the fuel cost analysis is already baked into the forecast, it would be interesting to draw that out in the conclusions more clearly. 

A: MISO should incorporate state EV and building electrification policies[3]

MISO should also incorporate state policy into EV and commercial and residential forecasts. The whitepaper indicates that MISO uses historical data by states to forecast EV loads, but state policies might force higher levels of EV adoption beyond what historical data might indicate. Likewise, state policies for building electrification might drive electrification beyond what MISO’s total cost of ownership methodology would suggest.

Q: What additional collaboration opportunities could strengthen alignment between MISO and stakeholders?

A: The Environmental Sector would like to partner on this effort

Stakeholder input should be broader than LSE engagement. The Environmental Sector is eager to be a resource to MISO, particularly to anticipate policy change and decarbonization trends. Many of the Organizations that make up the Sector have place-based colleagues who are active in state and city-level policymaking and it would be great to establish a communication path to pass those insights along to the MISO team. Give us a ring!

A: Direct input from large load customers, the Organization of MISO States (OMS), and active dialogue with survey participants would strengthen the forecast

MISO's LTLF team should engage large load customers directly, if they have not already. It is clear from the report that the unknowns regarding large load integration (e.g., plans for on site generation as BTMG or ZGIA) have the most consequential impact to this forecast.

MISO should also work with OMS and states in calibrating projections for data center growth rates post 2030. State policy can make a significant impact on where and how MISO should forecast data center growth in the medium to long term, and insights from states should be considered alongside third party data.

The Sector supports the pilot survey effort and urges MISO and OMS to continue to encourage participation from the LSEs. If not already part of the process, MISO's LTLF team should have an active dialogue with LSEs about what trends they are seeing that the survey may not be able to capture (e.g., interim milestones toward policy goals, local opposition to data center development, delayed timelines, etc.)

A: More clarity is needed on how the LTLF will inform the Futures

The April 13 presentation clarified that the 2026 LTLF will not directly inform any Futures cycle but that the next LTLF is likely to inform the next Futures iteration. We request clarity / formalization on how MISO adapts the LTLF for use in the Futures process and how it works in tandem with the parallel Member Survey, Scenario Definition Updates, and Assumptions Refresh & Model Build per the 2025 Futures Redesign Project Status & Schedule slide 5. 

Q: What aspects of MISO’s uncertainty approach need greater transparency, including any areas where forecasts may consistently over- or under-estimate expected conditions?

A: A long-range forecast requires a more durable Federal climate policy posture.

The whiplash in Federal climate policy creates investment instability and long-term uncertainty - and makes MISO’s Long-Term Load Forecast incredibly challenging. The Sector appreciates the dilemma of anticipating Federal policy beyond the current administration, but the years beyond the current administration represent 85% of this long-term forecast and requires a more durable planning posture. For MISO to assume the current pro-emissions policy position, a clear outlier on the world stage, as the status quo for the next 20 years is insufficient. The Sector recommends MISO look to peer countries across the globe for reasonable decarbonization commitments and ambition to anticipate electrification and fuel-switching trends that are very likely to be reinvigorated at the Federal level within the timeframe of this forecast.

The LTLF relies heavily on assumptions regarding how the OBBBA will drive home electrification, EV adoption, and distributed residential and commercial PV through tax credits. However, these policies are subject to significant political uncertainty over the forecast’s 20-year horizon. Since future administrations may change or even repeal these incentives (much like the reversal of the IRA) we recommend that MISO include a "policy sensitivity" analysis. This would demonstrate how the forecast could change under different political and regulatory scenarios.

Q: Are there any timing considerations we should be aware of as MISO establishes a predictable cadence for releasing load forecasts?

A: The Sector recommends an annual cadence

The Environmental Sector supports MISO establishing a predictable annual LTLF release cadence. The LTLF, OMS-MISO Survey, and PRA all occur in close proximity. MISO should consider ways to sequence or integrate these important milestones to ensure they inform one another rather than being isolated or siloed events.