Utility Capital Planning: Seven Insights from Industry Leaders

In September 2026, utility asset managers, system planners and engineers gathered in St. Louis, Missouri, for the CEATI Asset Management & Innovation Conference. CEATI (Center for Energy Advancement through Technological Innovation) hosts this annual event so that utilities can share practical experience in asset management. Our team attended to learn from peers and to compare what we heard with what we see in utility capital planning every day.

Several industry leaders took the stage, including Duke Energy, PG&E, Ameren Illinois, Hydro Ottawa and Swissgrid, alongside consultants and researchers who work with utilities. Their sessions covered transmission planning, asset health models, wildfire risk, hydropower operations, regulatory filings and artificial intelligence (AI). Yet one shift connected them. Utilities are no longer only asking which assets to replace. Instead, they ask a broader question: among everything we could invest in, what should we do, when, and why?

In short, utility capital planning now calls for plans that connect aging assets with future grid needs, show uncertainty openly and hold up under regulatory review. Better data and AI help, but only when every decision stays traceable. Here are seven insights behind that conclusion.

Capital budget optimization for utilities

Most utilities build their capital plan from a long list of candidate projects. Each project comes from a different team, for a different reason, and competes for the same budget. Several presentations showed that this list now far exceeds the money available.

These needs also come from very different drivers: reliability, load growth, regulatory obligations, modernization, and information and operational technology (IT/OT). A pole replacement program and a new substation therefore compete for the same dollars.

Ranking projects is not enough to settle that competition. When teams fund the highest-scoring projects until the money runs out, they can miss part of the value available. By contrast, comparing combinations of projects under real constraints produces a stronger investment portfolio. This is now a core challenge of utility capital planning, and one we explored in The budget is fixed. The risk isn’t.

Planning asset renewal and load growth together

Many organizations still treat asset renewal and grid growth as separate disciplines. On one side, asset teams track condition, deterioration and risk. On the other, system planners study load forecasts, capacity and new infrastructure.

Duke Energy, one of the largest electric utilities in the United States, delivered a keynote on building resilient energy systems for the grid of tomorrow. It showed the value of bringing renewal and growth together. Every replacement decision is also an opportunity to reduce risk, add capacity and prepare for future operating conditions. The speaker therefore invited utilities to test each major asset decision against one question: does it reduce today’s reliability risk while enabling tomorrow’s grid?

In Québec, Hydro-Québec made a similar point. Under its Action Plan 2035, the utility is adding more than 10,000 MW of new wind capacity to meet growing demand. Its grid planning team studied where that capacity should go and which transmission reinforcements it would require. The model showed that planning both together lowers total system cost. In the scenarios studied, reinforcing the grid too early raised average energy costs by 10 to 20%.

For utility capital planning, the practical question is simple. Why replace an aging transformer with an identical unit if the load forecast shows the substation will need far more capacity within a few years? Integrated planning of renewal and growth avoids that kind of decision. For a closer look at large new loads, read Powering growth without overburdening ratepayers.

Risk-based asset prioritization

Similarly, Swissgrid, the operator of Switzerland’s transmission grid, presented how it turns asset data into investment decisions. Its approach links its asset registry, data governance, an asset health index and a risk criticality index. Combined, condition and criticality produce an urgency rating that helps teams focus on the most critical assets.

This approach answers an essential question: where should we act? However, two more questions matter just as much. When should we intervene, and which lifecycle strategy creates the most value over time?

The answer is not always to replace. ATCO Electric, an Alberta utility facing more severe wildfire seasons, presented how it ranks poles by wildfire risk. Its key point: each risk driver calls for its own mitigation, from covered conductor and vegetation management to pole replacement or undergrounding.

Grant County Public Utility District (Grant PUD), which operates hydroelectric plants in Washington State, modeled controls such as keeping a spare transformer. It then compared their value with asset replacements. Likewise, a Penn State doctoral candidate presented a hydropower modernization case from Panama. The project weighed three options for aging control systems: keep the legacy system, retrofit part of it or modernize it fully.

A modern plan therefore compares complete lifecycle strategies: inspect, monitor, maintain, refurbish, defer, replace or expand. Learn more in What is asset lifecycle management?

Uncertainty in utility capital planning

Many utilities still assess risk with scoring scales and colored risk matrices. Grant PUD explained why it moved away from that approach toward a quantitative model. Verbal scales such as “likely” or “very unlikely” can create an illusion of agreement, because people attach very different probabilities to the same label. In addition, risk matrices make it hard to compare the return of different investments.

Asset life is also slow to observe. Failures can take decades to occur, so feedback arrives late. In that context, probabilistic methods such as Monte Carlo simulation and Markov models improve forecasts.

Chelan County PUD is a public utility in North Central Washington that provides electricity, water, wastewater and fiber services. It applied a probabilistic model to its underground cables to answer two questions. First, how long will the current budget take to reach a reliability target? Second, what is the chance of a worst-case scenario on a given circuit?

Our takeaway is that a probabilistic model and a clear decision do not conflict. Simulation describes the range of possible futures. Explicit rules, constraints and objectives then lead to a transparent, repeatable decision. We cover this balance in Reliability and resilience in electrical network planning.

Defensible utility capital planning

Regulatory scrutiny came up in session after session. Ameren Illinois, which serves central and southern Illinois, filed its first multi-year grid plan with the Illinois Commerce Commission, the state regulator, in 2023. The commission rejected it later that year. Among the reasons, the commission found that the plan did not quantify customer benefits enough and lacked a robust analysis of climate risk.

The utility responded with a cost-effectiveness framework and climate projections at the asset level. After a year of proceedings, the commission approved the refiled plan.

Duke Energy described a similar effort for proactive transmission upgrades. Its multi-value framework, accepted by the Federal Energy Regulatory Commission (FERC), counts reliability, resilience and congestion relief as part of the value case.

In addition, condition assessment specialists from the United Kingdom and the United States summed up the new expectation. Regulators no longer only ask whether a process exists. They also ask what the utility knew, what it could reasonably have known and what it did about it. Consequently, doing nothing is no longer a neutral choice.

For utility capital planning, regulatory traceability has become a real requirement: from engineering assumptions to an investment plan that holds up under review. It is the question behind every infrastructure budget. Our use case on transparent decision-making for regulators shows how comparing investment scenarios supports a stronger rate case.

Decisions from imperfect asset data

Utilities do not lack data. The challenge is that it sits across geographic information systems (GIS), enterprise asset management (EAM), inspection systems, SCADA, engineering models, spreadsheets and financial systems.

Several sessions showed that waiting for perfect data is not the answer. Chelan County PUD had good data for its underground cables but limited data for its overhead conductors. Even so, it built a useful risk prioritization by combining inspection reports, work orders, outage records and GIS layers. Its advice to peers was simply to start now.

Dominion Energy Virginia also started small. As its drone inspection program grew, the utility used AI on imagery it already had. The analysis showed that its records on bird deterrents barely matched what the images revealed, so the utility replaced those records with the verified results. Meanwhile, Swissgrid noted that its health index process led to more data collection and better data quality, because results now feed real decisions.

Other organizations measure progress directly. PG&E, which serves about 16 million people in northern and central California, is extending ISO 55001 asset management to its IT organization. That team tracks the completeness and accuracy of its asset records and certifies them twice a year. Another session focused on the buildings that house critical equipment. It proposed a decision readiness index, which measures the share of assets with everything a decision needs. That includes complete attributes, a current condition assessment, a risk score and a replacement value.

The lesson: start with the data you have, make the uncertainty visible and improve the model as data quality improves. Each system plays its own role, as we explain in EAM, CMMS and DSS: modern asset management layers.

Governing AI in asset management

AI appeared in most presentations, from drone inspections to hydropower dispatch. The concerns were consistent: governance, traceability, accountability and human oversight.

A session on AI in asset management, led by an international asset management consultancy, put it plainly: the goal is not more AI, it is better asset management. Human involvement should match the consequence of an error. AI can classify routine work orders, for instance, but setting risk appetite and approving asset strategies remain human decisions.

The examples followed the same pattern. In a session on hydropower operations, the planning system prepares dispatch options and the dispatcher approves the response, with a record of each decision. In transmission planning, Manitoba Hydro International helped develop an AI assistant that turns network studies into business cases. This Manitoba Hydro subsidiary delivers power system studies in more than 120 countries. Every assumption the assistant proposes must show its source before a planner approves it.

For capital planning, this means a glass box, not a black box. Every recommendation should be traceable from asset data to assumption, scenario and final investment decision. A transparent model also keeps expert knowledge in the organization, a growing concern as experienced staff retire.

Next steps for utility capital planning

Hydro Ottawa, the electricity distributor serving Canada’s capital, presented its roadmap for transforming enterprise asset management (EAM). It showed where the industry is heading. Its target ecosystem includes EAM, enterprise resource planning (ERP) and asset investment planning (AIP), each with a clear role. The AIP capabilities it expects are telling. They include less manual work, integrated data and the use of asset condition. They also include risk-based priorities, optimized timing and defensible justification for funding and rate applications.

Taken together, CEATI 2026 points to one need. Utilities need a decision layer that connects the condition of today’s assets with tomorrow’s grid, under real budget constraints and uncertainty.

That is the problem Direxyon’s solutions are built to solve. Direxyon Enterprise lets planners build custom asset models with decision trees and degradation curves, run Monte Carlo simulations and optimize the investment portfolio. As a result, teams can compare lifecycle strategies and turn the results into an actionable project list. AI completes missing asset data, so planning can start with the information already available. Direxyon works as the planning layer above EAM, GIS and asset health systems, rather than replacing them.

For electric utilities, our power utilities solution connects technical performance with financial outcomes, so each investment can be weighed against reliability, risk and long-term cost. Our whitepaper Aiming True: Capital investment for a reliable electrical network explores the subject in more depth.

CEATI 2026 confirmed that these challenges are shared across utilities, countries and network types. We look forward to continuing the conversation. Book a demo to see how scenario-based planning can support your next capital plan.h with questions first. 

Frequently Asked Questions