Login

Medium and Long-Term Load Forecast Workshop Load Forecast White Paper (20241218)

Item Expired
Related Entity(s):
Planning Advisory Committee (PAC)
Topic(s):
Transmission Planning

In the Decemeber 18, 2024 Medium and Long-Term Forecast Workshop stakeholders were invited to submit feedback the Load Forecast presented in the white paper and at the workshop by January 15, 2025. We request feedback on the long-term magnitude of the load drivers (e.g. Data Centers, Building Electrification), the growth rate of the drivers, and the range across load trajectories. We request that feedback focus on suggestions that have a material impact on aggregate volumetric or peak load. MISO will consider whether feedback should be reflected in our upcoming Futures update or a subsequent load forecast update.

Submitted Feedback

Issue:  Data center hourly load profiles to be used in the medium and long-term load forecasting process

 

MidAmerican has observed that data center facilities have different hourly load profiles depending upon where they are in the ramping process.  As a given data center facility is in the process of ramping up to its full capacity, its annual and monthly load factor will fluctuate due to testing, learning and load growth.  During the ramping process, the data center facility’s annual/monthly load factors will not be at the 90% level that is widely noted.  Once that given data center facility has reached its full capacity, its annual and monthly load factors will be in the 90% plus range.  MidAmerican recommends that MISO thoroughly research data center hourly load profiles and make use of different data center hourly load profiles based on their progress in ramping up to full capacity.

 

The 70% to 85% load factor referenced in the workshop is likely some average that does not take into account the ramping process of the data center facilities.  In addition, the data center facilities have not shown to be nearly as weather-responsive as was implied during the call. 

DTE Electric Feedback

DTE Electric (DTE) appreciates the opportunity to provide feedback on the MISO Long-term forecast approach. We agree that several industry developments could lead to significantly higher load growth in the next 5-20 years compared to the past two decades. However, we believe that MISO’s updated load forecasts may be overly optimistic, particularly when it comes to referencing policy adoption as a driver for vehicle and building electrification. While policy drivers can be a powerful force in driving adoption of new technologies, we believe that market forces and customer behaviors must have an equal influence on the long-term outlook on new technology adoption.

From our perspective in Michigan (LRZ 7), the MISO forecasts appear overly aggressive. We are concerned that this higher forecast could lead to recommendations and implementations of costly grid upgrades that could result in stranded assets. We recommend exercising caution when relying on policy initiatives to drive new technology adoption forecasts.

DTE Electric Recommendations:

  1. The higher growth projections are based in part on the IRA, IIJA, and the CHIPS act enacted during the Biden Administration. Federal and State Policies can change as administrations change. The IRA tax credits, IIJA and CHIPS act are all at risk for modification as administrations change, potentially impacting the long-term outlook for new technology adoption.   We urge caution in adopting forecasts based solely on policy. A flexible plan that takes into account the uncertainty of the federal policy is recommended.
  2. The latest MISO forecasts are higher than what we currently see playing out in Michigan, especially in the areas of EVs and Building Electrification. The incentive programs that are referred to as an underlying driver for both EV and Building electrification growth are small (0.05% for MiHER and the unapproved MI vehicle incentives combined do not have a material impact on load, short or long-term).
  3. It is not clear from the white paper whether you are considering regional views for building electrification. We recommend regional or zonal outlooks within the MISO footprint for building electrification technologies. DTE/Michigan is not experiencing a fundamental shift in the adoption of building electrification technologies, especially in communities with existing fossil fuel infrastructure.  Absent policy measures that prohibit the use of fossil fuel heating or cooking, we do not foresee a consumer driven shift in behaviors toward electrified heating at the magnitude the MISO forecasts suggest.
  4. EV manufacturers are pulling back on their initial EV adoption forecasts.  Adoption rates differ significantly across the country (California 27%, Michigan 5%) and by region.  We also recommend regional outlooks for vehicle electrification adoption forecasts and recommend that the EIA forecast for EV adoption be added as a low-end estimate.  (Comparing Figure 15 (MISO Current Trajectory Energy Consumption (TWh) by Vehicle type), MISO is at 28% CAGR total from 2023 to 2040. The comparable DTE Electric forecast is 23%.)
    1. Comparing the DTE Electric charging profile to MISO’s profile shown in Figure 16. The DTE profile, based on DTE customer-specific AMI data, indicates that incentives have shifted charging to later in the evening compared to the MISO home profile.  We anticipate that the impact on peak load will be minimized in the long term as EV owners naturally gravitate towards off-peak charging.  We recommend MISO investigate this trend and incorporate this data into their peak modeling efforts.
  5. Data Center load, as indicated in the MISO white paper, indicates that public announcements were used for data center growth projections.  DTE is concerned that this approach to forecasting data center load could be overly optimistic.  As these projects may not materialize, DTE is concerned about potential double counting in load estimates, as projects may be considering multiple utilities or states. Additionally, DTE recommends that future efficiency gains be considered in future data center load projections.
  6. Using overly optimistic assumptions for emerging technology adoption poses a risk. In Figure 1, the stakeholder input aligns most closely with the low end. We recommend aligning the stakeholder input with the midpoint and lowering the low forecast further. The low forecast should depict a scenario where most policy drivers do not materialize. The stakeholder forecasts should inform the middle "current trajectory." We suggest using low-end estimates similar to those issued by the EIA.

MPSC staff seeks clarification on the treatment of Distributed Energy Resources (DERs) in MISO’s Medium and Long-Term Load Forecast as explained in the Whitepaper.

Page 4 refers to DERs as capacity, page 7 portrays DERs as a driver of energy demand increase, and page 8 discusses the contribution of DERs.

In the refences contained above, the language used to characterize the effect of DERs on net system load is unclear. In the absence of energy storage, DERs are conventionally thought of as having a net reduction to system load as they reduce the amount of power needed from the bulk power system. Energy storage, depending on when it is charging/discharging, could result in a net addition to system load, depending on the timeframe being considered. MPSC staff would prefer a consistent reference to the impact of DERs in the context of the forecast, or at the very least, an explanation as to why DERs are treated differently.

MISO should provide LRZ specific figures for each driver of load growth.

In MISO’s efforts to improve the granularity and industry-specific segmentation, if MISO is able to provide LRZ specific breakdowns of the forecasted impacts of each sector (e.g. building electrification, data centers, electric vehicles, DERs) specified in the Whitepaper and all future iterations, this would prove extremely useful for forecast analysts residing in each zone.

These comments and questions are submitted on behalf of the Staff of the Louisiana Public Service Commission ("LPSC Staff"), and they do not necessarily reflect the views of the Louisiana Commission.  The LPSC Staff position is that the local load serving entities ("LSEs") are the best source of load forecast information because they have better access to the data and other information needed to make load forecast projections.  The local LSE information should form the backbone/basis for MISO's medium and long term load forecasts.  In addition, the whitepaper should provide detailed information on the issues, proposed solutions and methodologies.  This whitepaper essentially informs the reader about what MISO has decided to do, but it provides woefully inadequate data-driven support for MISO's conclusions, particularly for the projected levels and specific locations of load growth in each LRZ.  In addition, MISO's top-down load forecasting proposal fails to respect principles that MISO previously followed, including:

  • LSEs have more insight and wisdom into long term load growth.
  • LSEs have access to the best data and are more knowledgeable on their own load.
  • LSEs have the best information upon which to base reasonable load growth projections for medium and long term purposes.

The LPSC Staff does not object to MISO gathering additional cost-effective information to supplement and check LSE forecasts, but that additional information should not be the primary basis for forecasting decisions that are being used to support billions of dollars in transmission investments. A top-down forecasting "whiff" could result in costly investments in excess transmission and/or transmission located in the wrong locations.

 

As a result, LPSC Staff offers the following feedback and seeks the following information related to MISO’s Long Term Load Forecast whitepaper dated December 2024:

 

  1. On Page #3, MISO assigns stakeholder submitted load forecasts to the “Low Trajectory” scenario.  Please explain why stakeholder submitted load forecasts are not the foundation for the “Current Trajectory” scenario as stakeholder (assuming these are LSE load forecasts) submitted data will best represent current expectations of load growth.
  2. Previous Futures Load Forecasts began with the development of the EGEAS-Ready Coincident Peak (CP) Demand and Energy Forecasts for each Future and with MISO’s LSE’s 20-year demand and energy forecasts and ended with the application of the various Future-driven assumptions, creating Future- and year-specific forecasts. The load growth was based upon reaching a targeted energy increase 20 years out.  Please explain in detail how MISO intends to change from this historical methodology.
  3. On Page #4, MISO notes the rise of rooftop solar, energy storage and other DERs is expected to contribute 69-78 "TWh" of capacity to MISO’s system by 2044.  Is this reference correct?  Please provide the details by year and by region of MISO’s expectations for DER capacity contributions through 2044.
  4. On Page #5, Figure #2 references CAGR with what appears to be high values.  Please advise if the CAGR values denoted in Figure #2 are correct.
  5. On Page #5, Figure #2 MISO denotes a “Level of Certainty” for each sector load growth.  What is the foundation for the “Level of Certainty” values?  Is this MISO’s opinion?  Please provide supporting documentation affirming the assigned “Level of Certainty” values.
  6. On Page #6, MISO states MISO plans to incorporate its updated forecast approach to its future planning assessments.  Please advise if MISO intends to use LSE submitted load forecasts as the foundation for the Futures.  Does MISO intend to work closely with LSEs and State Regulators to incorporate its load growth projections into the LSE load growth forecasts?  If not, why not?  LPSC Staff notes that State Regulators and LSEs are closely attuned to anticipated load growth (more so than MISO) in their native LRZs.  In view of the impact on ratepayers, please confirm that State Regulators and the LSEs will be the final arbiters on load growth projections for each LBA and associated LRZ.
  7. On Page #7 and throughout the whitepaper, MISO references federal programs as expected drivers for load growth in the various Sectors.  However, these federal programs have designated expirations, income limits for federal tax credits and other limitations that could substantially diminish their impact going forward and most notably after 2032.  How did MISO account for these limitations and potential expirations beginning in the early 2030s?
  8. On Page #7 and throughout the whitepaper, MISO references State programs in a limited number of MISO States as contributing to load growth for various Sectors.  However, these State programs have a number of limitations that impact their long-term value.  How did MISO account for these limitations?  How did MISO account for the lack of such programs in the majority of MISO States?
  9. On Page #7, MISO notes anticipated Data Center energy demand increases of 149-241 TWh by 2044.  Please provide relevant documentation describing the analysis MISO used to arrive at these TWh values.  LPSC Staff hereby expresses its deep concern with MISO independently engaging consultants and incurring other expenses to determine potential Data Center sites when this type of information is best available at the LSE level.  The incurrence of these consultant costs and other expenses is especially concerning in view of MISO’s rapidly increasing costs that ratepayers must bear.
  10. On Page #7, MISO references expected growth in EV adoption after 2030 with MISO basing these expectations on federal and State programs as key drivers.  How did MISO account for potential expirations and other program limitations, most notably after 2032?
  11. On Page #7, MISO references EV adoption as a key driver for increases in energy demand.  Did MISO account for the impacts of potential rare earth availability limitations and potentially rapid price increases in rare earth prices that will negatively impact EV costs and resulting adoption?  If so, please provide a narrative on how MISO accounted for these impacts.
  12. On Page #8, MISO references the need for continued long term federal subsidies to support green hydrogen.  Please provide a narrative explaining how MISO will treat load growth attributed to green hydrogen should the needed long term federal subsidies not exist.
  13. On Page #8, MISO references DER as contributing 69-78 TWhs to MISO’s demand.  Please explain how MISO feels DER will contribute to MISO’s demand.
  14. On Page #8, MISO states MISO must also account for potential economic headwinds and recessionary pressures which could temper the rate of load growth from these industry developments.  Please provide a narrative as to how MISO will account for factors that could temper the rate of load growth.
  15. On Page #9, MISO references stakeholder outreach and other inputs.  Please provide a narrative on what stakeholders were contacted, what inputs they provided, along with what other inputs were utilized.
  16. On Page #9, MISO states Ensure alignment with evolving regulatory and Tariff needs.  Please provide a narrative describing this alignment.
  17. On Page #9 and throughout the whitepaper MISO references “progressive forecast methodologies.”  Please describe in detail what MISO means by its reference to “progressive forecast methodologies.”
  18. On Page #13, MISO describes where it anticipates load growth to occur, and Sectors that will contribute to this load growth.  Does MISO intend to work with the respective State Regulators and LSEs for the respective LRZs to gain their approval or does MISO intend to apply these load growth energy demands without verification from the entities with the best source of information on these issues?
  19.  On Page #17, MISO states MISO is poised for rapid data center demand growth due to low energy costs and favorable generation and transmission capacity relative to other regions.  Please explain why data centers are heavily concentrated in other non-MISO regions (i.e. PJM East and ERCOT) if MISO has these favorable metrics.  In view of MISO publications depicting a negative outlook for sufficient generation, why does MISO use the term “favorable generation?”  How did MISO determine it has “favorable transmission capacity?”
  20. On Page #18, MISO begins its discussion on Data Centers and expected load growth.  Please see Comment #9 above referencing LPSC Staff’s deep concern with MISO’s process and associated costs with Data Center load growth especially when needed information is generally available from the LSEs at a much lower cost.  MISO states Accurately forecasting data center growth presents significant challenges.  Please provide a narrative on the methodologies to be utilized accounting for these challenges.
  21. On Page #20, MISO states Local adjustments are made at the LRZ and LBA levels based on historical data.  Please provide documentation detailing the adjustments to LRZ 09 and adjustments to LBAs specific to Louisiana.
  22. On Page #25, MISO states the northern states in MISO’s footprint are especially positioned for accelerated manufacturing growth due to transmission capacity and low-cost resources.  Please advise how MISO determined the northern states in MISO’s footprint have transmission capacity available and low-cost resources.
  23. On Page #29, MISO discusses DER contributions.  Please provide a chart depicting MISO’s expectations for annual DER MW additions for each year, for each zone through 2044.

 

TO:                       MIDCONTINENT INDEPENDENT SYSTEM OPERATOR

FROM:                 THE ENTERGY OPERATING COMPANIES

SUBJECT:             MEDIUM AND LONG-TERM LOAD FORECAST WORKSHOP LOAD FORECAST WHITE PAPER (20241218)

DATE:                   JANUARY 15, 2025

 

The Entergy Operating Companies (Entergy) appreciate the opportunity to comment on the Medium and Long-Term Load Forecast proposal defined in the MISO Long Term Load Forecast Whitepaper and as presented at the workshop on December 18, 2024. Entergy offers these comments on long-term growth projections with a focus on MISO South and the Local Resource Zones (LRZ) in MISO South.

The proposed MISO TWh growth from approximately 640 TWh in 2025 to 921 - 1,225 TWh in 2044 appears reasonable when compared with Entergy’s long-term growth expectations. We believe the growth projections at the LRZ level are generally reasonable for the primary growth categories MISO expects to impact MISO South.  As demonstrated by the significant changes from Series 1 to the current proposal, load forecast factors can change quickly, and MISO will need to work regularly with local stakeholders to ensure that load growth assumptions reflect the latest information and continue to be appropriate – and that these data are refreshed prior to undertaking new planning assessments.

LRZ9 has the widest range of uncertainty of any zone, with the high scenario being approximately 300% of the low scenario. The wide range is likely due to the "high uncertainty" rating of the primary drivers of growth MISO predicts for LRZ9. Entergy recommends that MISO consider reducing the upper bound of the range to make it more consistent with the other LRZs projected to experience high growth. Additionally, LRZ9 has the highest growth potential of any LRZ at 117 TWh in the High scenario. This may be reasonable given the size of LRZ9 relative to the other zones and the potential growth drivers. However, the growth drivers that MISO has identified for LRZ9 are highly specific to several regions within LRZ9.  For example, for industrial load growth, it would not be reasonable to assume that growth occurs in locations within LRZ9 in which industrial load is not anticipated to occur.  An approach more granular than the LRZ level will likely be needed when applying the load forecast to LRZ9 to ensure reasonable planning outcomes.

AI and Data Center

Entergy recognizes the importance of including data center growth in long-term planning and supports MISO efforts in this area.  Although MISO did not identify this growth driver as a part of their long-term load forecast in MISO South, we offer these comments in case the forecast changes.  Entergy sees several challenges MISO will need to address when including large point-specific loads beyond the 10-year planning horizon.

  • Solutions that accommodate large, point-specific loads are typically costly and depend significantly on the customer's interconnection point. Including these large, point-specific loads in planning models can result in transmission solutions that depend heavily on the presence and location of the load to achieve estimated benefits. In other words, if the precise location of the load changes even slightly, the benefits of the project may not be realized.  The uncertainty and sensitivity regarding the precise location of forecasted large customer loads should be considered in MISO’s planning processes if they are to be included in the models.
  • Incorporating speculative large loads into long-term planning models may result in shifting network upgrade costs from prospective customers to existing ones. This shift could affect customers who are not within the same state and may not benefit from the transmission investment in the same manner the interconnection customers will.  MISO should carefully consider the weight any single load decision has on the benefits of a network upgrade.

Green Hydrogen, New Industry Development and Reshoring

MISO identified Green Hydrogen, New Industrial Development, and Reshoring as key growth drivers in MISO South. We agree that these areas show significant potential along the Gulf Coast. However, the magnitude and type of load growth will vary significantly across different parts of LRZ9, so we urge MISO to collaborate with local stakeholders with detailed knowledge of expected types and locations of customer load projects when applying these assumptions to MISO’s models.

Electric Vehicles

MISO has increased their EV assumption and raised the certainty level compared to Series 1 Futures. Entergy also anticipates that growth will accelerate in the 2030s. It is expected that EV growth in MISO South will be gradual, primarily impacting densely populated urban and suburban areas. We recommend that MISO focus the projected EV growth in MISO South in these specific areas.

Building Electrification

MISO indicated that their growth expectations for building electrification have been reduced compared to the Series 1 Futures. MISO South currently lacks any formal and significant policies anticipated to drive substantial building electrification investments. While Entergy expects some residential and commercial electrification to occur, Entergy does not foresee that factor significantly affecting long-term load forecasts currently.

Comments: 

  1. Is the forecast of load or obligation?  Currently, MISO’s determines obligation based on the MISO coincident peak load. The coincidence factor between LBA peak and MISO CP equates forecasted load to forecasted obligation. MISO’s proposals to reassign obligation based on historical performance during RA hours, results in a difference between a peak forecast and a forecast of obligations. An LBA could have a forecast with no change in load, however due to changes to the grid outside their control and outside their territory, the LBA’s obligations can change. The implications of this distinction become more acute when examining the forecast in greater granularity. 

  1. Why doesn’t MISO match load to obligation? MISO discounts resources based on probability of availability during RA hours, but does not share what conditions determine the RA hours. However, on the forecast, there is no discounting of the forecast. By definition, load at the peak will be greater than or equal to the load at the RA hours. To match load with resources, the forecast should be discounted to reflect expected load during the RA hours and not simply reflect the peak hour. 

  1. MISO needs to clarify “reliability” when considering an hourly forecast. The MISO long term forecast is an hourly forecast rather than a seasonal hourly peak value. This raises a question of what constitutes reliable coverage of hourly forecasted load. Is the goal to cover 110% of the load for 95% of the hours, or 103% of the load for 100% of the hours?  

  1. Recognize that forecasting obligation hours is less deterministic (aka. more dependent on MISO data and LOLE calculations) than when the obligation hour was assumed to be the MISO peak. Historically, residential customers were variable, but predictable based on weather and could be included in a peak model. Industrial customers that were variable, but not weather sensitive, could be statistically incorporated into a peak model. Because of MISO’s new method, statistical models are less certain (we are dependent on MISO to determine the obligation hours). Going forward MISO needs to come up with a way for LSEs to calculate obligations through a process that is reliable and repeatable.  

  1. More transparency is needed in MISO's load growth forecasts. MISO has admitted that the forecasts are increasingly less econometrically derived and are based on a set of fundamental assumptions informed by expert opinions, secondary research, and information from LSEs regarding data centers. MISO should consider sharing the underlying assumptions with LSEs to improve consistency across their forecasts. 

Examples: 

    • Building Electrification: Why is widespread adoption delayed until 2030 despite the availability of heat pump technology, especially in the commercial and industrial sectors? Some customers in our service area have already transitioned to electric heating. 

    • EV Adoption: How might the pace of electric vehicle adoption slow under the second Trump administration? 

    • Data Centers: 

      • Can MISO provide more details on the 41% attrition rate? Specifically, how was it determined, how is it being applied, and is it updated based on actual load studies? 

      • Is MISO ensuring that data center projects are not being double-counted by different Load Serving Entities (LSEs)? 

    • Industrial Load and On-shoring: What impact could tariffs, a reduced labor supply, and resulting inflation under the second Trump administration have on new industrial load and on-shoring efforts? 

  1. MISO forecasts should also address customer costs in the forecast. Additional load will add pressure to the electric grid. As those investments are made, consumer costs may increase. Ignoring this potential increase overstates future loads.  

  1. MISO should be transparent in growth opportunities. In the presentation MISO highlighted  that data centers are currently locating in Iowa, Minnesota, and Indiana because these locations currently offer the lowest costs. At what point does MISO expand to make additional territories available? Where would that expansion occur? How much load growth can IA/MN/IN add before they are also constrained. 

  1. Address perceived bias. MISO’s stated primary concern is reliability. An aggressive forecast creates a perception that MISO must take action to relieve a perceived threat. This creates a perception that MISO may make aggressive assumptions in the forecast process. To address this perception, MISO should be overly transparent in making their process and results available. Additionally, when making assumptions, MISO should choose conservative assumptions. The downside of not predicting long term growth soon enough is higher prices due to a loss of flexibility, not the collapse of the grid. 

  1. Make forecast, including support, available to LSE. When LSEs provide long term forecasts in regulatory settings, they often need to provide support for that forecast. Since MISO is creating a long-term LSE level forecast of obligation at the LSE level that will translate to financial obligations, MISO should not only provide that forecast, but also the resources necessary to support the forecast to the LSE. Examples of resources needed would include: 

Common documentation of the overall process. 

Assumptions and results specific to the LSE. 

A Q&A mechanism for the LSE to ask follow-up or scenario questions. 

A resource identified at MISO, should regulators require testimony on behalf of the LSE.  

 

Questions:  

  • Timing: Is the forecast done and MISO is reporting results, or is this an opportunity to have input to the process? 

  • Is the low forecast low enough? 

  • Does the MISO forecast incorporate any impacts due to the higher electric prices that would be associated with expanded generation or transmission assets to serve the additional load?  

  • How frequently does MISO intend to update the forecast and how does future planning change from as the forecast changes? 

  • Does the forecast documentation list the major risks to the forecast? 

  • Will MISO be updating the forecast on an annual basis? Which version of the forecast will feed to the LOLE model that has resource implications? 

  • In reassigning the MISO forecasted load back to the LRZ and LBA, what determines the allocation? Is it based on the percent of load during RA hours, during peak hours, based on historical calendar days? Why does MISO feel that allocation is the most appropriate? 

  • Does MISO anticipate LBA’s helping to mitigate future load increases or encouraging electrification by managing load through rate design, such as mandatory time-of-use rates, critical-peak pricing, or off-peak discounts? If so, what information is available to help design those rates. (Said differently, Does MISO promote any price structure other than cost + pricing based on real time prices?)  

Thank you for the opportunity to provide considerations for this issue. Accurately predicting future load growth is complex and requires a thorough understanding of the underlying data. While the alignment of MISO’s forecasts with initial GenAI driven demand is beneficial, it seems coincidental rather than predictive, warranting caution when predicting future trends. Paraphrasing Eisenhower “plans are worthless, but planning is everything,” which demonstrates the value of robust planning based in clarity and transparency.  

The absence of detailed data tables, quantified weighting factors, and clarity on third-party data sources makes it difficult to provide many comments of substance to address the magnitude or growth rate of the drivers. Access to this information can enable stakeholders to validate assumptions and lead to more informed discussions.

SECTION 1: LONG-TERM MAGNITUDE OF THE LOAD DRIVERS (E.G. DATA CENTERS, BUILDING ELECTRIFICATION)

The long-term magnitude of the load drivers includes several areas the deserve further investigation:

  1.  Data Center Growth: Data centers are a dominant driver of incremental load increases, with MISO estimating they will account for ~65% of growth by 2028. However, regional variability complicates broad comparisons. For example, a Bain study (1) anticipates data centers to account for 44% of national electric growth by 2028, but some areas like Georgia could likely see an even higher growth. Additionally, load forecasts should consider efficiency gains in AI hardware and the potential for an S-curve adoption which could moderate these projections. Further transparency into MISO’s methodology for data center growth in the MISO footprint, how technological improvements are accounted for, and how consumer adoption is factored in is critical to provide feedback on the magnitude of this driver.
  2. Interaction Between Assumptions: The interaction between drivers deserves further investigation. For example: building electrification is linked to the change in degree days and state & federal policy; Distributed Energy Resources (DERs) can potentially mitigate peaks through EVs and building electrification; macroeconomic load growth in the wake of GenAI and data centers may further increase peak load; state and federal policy can further drive or mitigate peak load; and the time and location of EV charging patterns can impact peak load.
  3. New Industry & Reshoring: In the conventional industry growth section, electrification of oil, gas, and chemical/petrochemical sectors is expected as a driving factor along the Gulf Coast. Regional load growth can also be impacted by electrification of agriculture (some of the most fertile land in the US is in the MISO footprint), electrification of manufacturing (green steel has significant potential and could raise energy consumption), and the chemical industry (which has potential to relocate as new technologies are developed). Data supporting these decisions will help stakeholders to better validate the magnitude of this driver.  
  4. Green Hydrogen: Although the possibilities of this technology are exciting, its adoption depends on an excess of renewable energy and additional obstacles outside of the grid and energy production. The DOE’s pathway document (2) estimates an additional 200 GW of renewable resources are needed by 2030 in addition to supplying other incremental load growth. As an example outside the grid, aviation will require changes for liquified storage, aircraft redesign, and updated regulations which, unfortunately, might suggest a slow adoption.

SECTION 2: GROWTH RATE OF THE DRIVERS

Transparency around the data and the weighting of the drivers is important to provide input on the growth rates.

  1. Data Transparency: Access to detailed data tables, quantified weighting factors, and clarity on third-party data sources is necessary for verifying the CAGR assigned, respective weighting for each primary driver, and how terms like “aggressive adoption” are quantitatively defined. Third-party data are often cited in the footnotes and inconsistently defined in the footnotes and body of the whitepaper. An understanding of how third-party data was collected, its limitations, and potential biases can provide further insight into how these trajectories were established. Finally, understanding whether this analysis is driven by MISO internal data or data provided by the LSEs could help clarify these trajectories.
  2. Technology Efficiency Gains: As stated in the previous section, this does not seem to include efficiency gains and an S-curve adoption of GenAI or EVs. AI hardware will become more efficient, especially as more competitors emerge and increase competition with NVIDIA. And conversely, this could encourage a faster adoption of the technology and raise incremental load growth. An understanding of how technological efficiencies are accounted for will help in understanding the growth rates.
  3. Population and Demographic Shifts: The Midwest is expecting a population decline of 2% by 2040, this should be accounted for in consumer behavior-based forecasting. (3)
  4. Data Center Trends: Some data centers are considering behind the meter generation to secure the level of reliability they need; this has potential to further reduce incremental Network Load growth due to data centers.

SECTION 3: RANGE ACROSS LOAD TRAJECTORIES

  1. Scenario Labeling: The current projections (Low, Current, and High Trajectory) and their definitions are appreciated, but a restructuring around the most divergent drivers could provide a more illustrative projection such as 1) Unconstrained supply chain with accelerated Data Center load growth, 2) current trajectory with technology efficiency enhancements, 3) imminent recession with reduction of federal incentives. Drivers with less weight can be used to create a sensitivity analysis for each of the main drivers to help increase clarity.

To enable robust stakeholder input, MISO should share additional information such as the underlying data, detailed data tables, quantified weighting factors, assumptions, and clarity on third-party data sources. This will help all stakeholders better understand the methodologies and provide more meaningful feedback to help address the challenges of future load growth.  

Linked Sources:

  1. https://www.bain.com/insights/utilities-must-reinvent-themselves-to-harness-the-ai-driven-data-center-boom/#
  2. https://liftoff.energy.gov/wp-content/uploads/2023/05/20230523-Pathways-to-Commercial-Liftoff-Clean-Hydrogen.pdf
  3. https://www.coopercenter.org/research/national-50-state-population-projections-2030-2040-2050

The OMS Resources and Transmission Planning Work Groups (OMS Work Groups) provide these comments to MISO in response to its Medium- and Long-Term Forecast Workshop and white paper. This feedback is from OMS work groups and does not represent a position of the OMS Board of Directors. 

The OMS TPWG and RWG appreciate the scope and effort put into MISO’s load forecasting white paper and workshop. 

1.) As MISO’s initial projections for substantial load growth are associated with a high degree of uncertainty related to data centers, we encourage MISO to provide stakeholders with as much information as possible about its assumptions, such as what its third-party consultants are advising and why forecasting future data center growth (i.e., new data centers that are not planned or announced) based on where data centers are currently planned or located is appropriate.

We would like MISO to further clarify its assumptions about DERs and energy efficiency: 

  • Is MISO contemplating DERs would increase load growth? At the workshop, MISO staff advised DERs may offset load, but the white paper is phrased to suggest DERs would increase load.  
  • Can MISO provide more details regarding its assumptions that energy-efficient appliances related to cooling would increase load? EPRI’s U.S. Energy Efficiency Potential Through 2040 report suggests reduction in demand from more efficient AC.
  • Could MISO provide more clarification on EV charging profile (Figure 16) by sharing a figure that shows how overall EV charging load changes by hour, rather than by location, as a percent of total charging load?
  • Can MISO specify if and how LSE forecasts were included in its forecast, or how they will be included in MISO’s projections in future versions of this study? The OMS Work Groups would appreciate further information on the various load forecasts MISO receives from LSEs and other external sources (e.g., Purdue's State Utility Forecasting Group, outside consultants) and how those forecasts are used in MISO's planning initiatives, operations, and markets. MISO should also demonstrate why LSE-submitted forecasts are less reliable or useful than other forecasts. 

2.) We request that MISO address the following concerns:

  • MISO’s locational pro-rata projections for data center growth could potentially have substantial consequences for benefit projections and ultimately cost-allocation. Therefore, proceeding with caution relative to the reliability of MISO’s projections is important for respecting the principles of cost-allocation and relatedly to protecting ratepayers and ensuring they are assigned costs appropriately. As it appears that the new forecasts are generally in range with the current series forecasts until after the next decade, as seen on Figure 6 of the white paper, we relatedly have concerns that there may be greater temporal uncertainty with forecasting data center load, which may also have cost allocation impacts.
  • As forecasts influence LRTP, MISO’s projections may create a “self-fulfilling prophecy” and disincentivize data centers locating in places that would otherwise make the most efficient use of existing transmission, existing generation, and/or seek to innovate to be more efficient.
  • MISO should also be cognizant of potential uncertainty regarding incentives or regulations for data centers at the state-level, as states are still reacting to data center growth and impacts.
  • We heard MISO staff suggest to some degree the concerns around uncertainty could be addressed through scenario development. We suggest that at a high-level, some of those discussions may need to take place in tandem with developing the forecasts to reassure stakeholders that forecasts with a high degree of uncertainty can be appropriately bounded or contextualized.

3.) To better grasp and contextualize the projected load growth, could MISO supply an incremental load growth bar charts with the sources broken out (i.e., one like Figure 10, 12, 14, 17, or 19) in units of net coincident peak load rather than TWh for comparison?

Further, as MISO’s new accreditation approach has trained stakeholders to think about risk hours rather than peak, it would also be ideal to demonstrate how the load growth impacts risk hours, even if just in aggregate (i.e., if Figure 6 of the white paper could be displayed in terms of risk hours) 

  • Based on the potential to incentivize off-peak (or off-risk) charging behavior via time of use rates for EVs and the degree to which green hydrogen production relies solely on renewables, some of the load growth sources may contribute differentially to peak as opposed to risk hours. 

4.) Through Order No. 1920-A, relevant state entities may seek from MISO the explorations of additional scenarios. Although MISO’s Order No. 1920-A stakeholder process has not formally begun and MISO is still in its initial stages of load forecasting, we hope that MISO is prepared to offer OMS the option to select additional future scenarios for analysis if needed. We mention this now as MISO discussed using scenarios to compensate for uncertainty in the workshop.

Related Issues

There are no Related Issues that match this criteria.

Related Materials

Supplemental Stakeholder Feedback

MISO Feedback Response