How Columbia’s Nuclear Uprate Cut $500M Without Losing a Watt

A nuclear uprate was supposed to cost roughly $1.2 billion. It now carries a price tag of about $700 million, and the scope did not shrink to get there.

What changed was the timing. Energy Northwest restructured when and how the work gets done, and that scheduling decision erased roughly half a billion dollars from the bill.

That project is the Columbia Generating Station uprate, scheduled to begin its first major outage in spring 2027. The engineering decisions being locked in now will determine whether the Pacific Northwest gains 162 megawatts of firm, carbon-free baseload capacity by 2031, or whether a $700 million plan drifts over schedule and budget.

Roughly 30 individual capital modifications need to be procured, sequenced, and executed across three refueling outages, all without taking the region’s largest single carbon-free power source offline for long.

Here is how a large-scale nuclear uprate actually gets executed: which components are replaced and when, why the work is split into phases, and what risks the project team is managing. Consider it practical literacy for anyone tracking U.S. nuclear capacity growth or Pacific Northwest energy infrastructure.

What the Columbia uprate is actually doing to the plant

The destination is a number: 120% of the plant’s originally licensed thermal power, or 3,988 MWt.

Reaching that figure is not a settings change. It is a re-engineering of how the plant converts heat into electricity, and that re-engineering cascades through valves, pumps, turbines, and the electrical distribution system.

Columbia currently runs at 106.7% of its original licensed thermal level. Pushing to 120% requires replacing or upgrading major components so the plant can safely handle higher steam flow, higher pressure, and higher output.

Here is the capacity math that the engineering unlocks:

• Current licensed thermal level: 106.7% of original
• Core extended power uprate (EPU) addition: approximately 162 MWe of net electrical output
• Efficiency improvements layered into the same outages: approximately 24 MWe
• Total anticipated increase: approximately 186 MWe
• Homes powered by the 162 MWe core addition alone: roughly 125,000

The project is classified as an extended power uprate because it exceeds a 7% output increase. That threshold matters: it triggers mandatory component replacements and a comprehensive Nuclear Regulatory Commission (NRC) licensing review, which separates this from simpler measurement-based uprates that squeeze out small gains without major hardware changes.

The NRC power uprate review process governs license amendments under 10 CFR 50.90-92, and the agency has approved 171 uprates since the 1970s, adding more than 8,000 MWe of capacity to the U.S. fleet.

What should reassure anyone assessing execution risk is that 120% thermal power is not a frontier.

A proven destination, not a first attempt Twelve of the 31 currently operating boiling water reactors in the United States have already reached the 120% thermal power threshold, a level standardised by General Electric. The NRC has approved more than 170 uprates across the industry since the 1970s.

That context tells you Columbia is navigating a known technical pathway rather than breaking new ground. For investors and energy observers, knowing that a dozen peer reactors have completed the same class of upgrade frames the project as de-risked at the design level, even though execution risk remains very real.

The Columbia uprate fits within a much larger picture of nuclear capacity expansion, where governments and utilities globally are committing to tripling installed nuclear output by 2050, a target that makes near-term uprates at existing plants economically and politically attractive relative to decade-long new builds.

The three-outage sequence: what gets replaced when and why

The uprate is split across three scheduled biennial refueling outages, and each one is a deliberate engineering chapter rather than an arbitrary slice of the work.

R-28, in spring 2027, is the foundation outage. Planned scope includes moisture separator reheater valve and nozzle upgrades, heater drain valve replacements, selected piping support improvements, in-vessel measurements, and turbine inspections. This outage prepares the plant for what follows rather than delivering the full output gain.

R-29, in 2029, is the pump configuration outage. Its signature modification replaces the current 3,000 horsepower booster pump motor with a 4,000 horsepower unit, a targeted capability step-up the plant needs before it can reach final operating conditions.

R-30, in 2031, is the completion outage. Reactor feedwater pumps are replaced with modern pump technology, alongside turbine generator upgrades, heat exchanger upgrades, and electrical distribution improvements. These are the modifications that finally bring Columbia to its 120% thermal power operating level.

The sequencing logic is dependency-driven. Each outage creates the physical and operational conditions the next set of modifications requires, which means a delay in R-28 does not just push one outage back; it compresses the entire chain toward 2031.

That is the key to reading schedule risk here. Because R-28 is a foundation outage rather than an implementation one, the 162 MWe gain does not arrive incrementally. It is unlocked at the end of R-30 in 2031.

Preparatory work already completed ahead of R-28

Energy Northwest has been de-risking R-28 for years, and several groundwork items are already done.

Laser scanning during the 2025 R-27 outage captured high-resolution spatial data inside key plant structures, feeding equipment replacement planning, outage sequencing, and radiation dose reduction. A concrete block shield wall installed in 2023 replaced a solid wall specifically to permit future installation of new moisture separator reheater components.

The turbine building has also received structural modifications, component staging, and access improvements to ready it for large component replacements. Hydraulic modelling confirmed that the existing condensate pumps can handle uprated flow with a motor upgrade to 1,300 horsepower, and a new condensate pump was installed in January 2025.

Design work for key valve upgrades is now 100% complete. Each of these milestones reduces the number of unknowns the project team carries into R-28, which is exactly where outage schedules tend to slip.

Why not do it all at once? The five-part logic behind phasing

If the plant is going to be shut down anyway, why not replace everything in one long outage? The answer is where this project’s economics live, and it explains how the cost estimate fell by roughly half a billion dollars without cutting a single component.

Phasing rests on five interlocking reasons:

1. Grid reliability. Completing upgrades during each scheduled refueling outage lets Columbia keep providing baseload power between outages. The Pacific Northwest grid never has to absorb one prolonged, concentrated plant shutdown.
2. Lifecycle alignment. Nearly half of the roughly 30 EPU capital projects overlap with maintenance and replacement work the plant needed regardless of the uprate. Folding uprate work into that already-scheduled activity eliminates the cost of separate outages.
3. Execution risk management. Spreading 30 major modifications across three outages keeps each one within typical durations. Staged execution also allows testing, validation, and correction between outages, so a single failure does not jeopardise the whole project.
4. Capital and financing efficiency. Phasing lets the Bonneville Power Administration (BPA) and Energy Northwest spread capital across fiscal years, smooth rate impacts on BPA customers, and align financing instruments such as lines of credit and bond issuance with staged construction.
5. Licensing strategy. The NRC’s projected review window runs from Q2 2028 through Q4 2030, and phased execution lets Energy Northwest implement early modifications and gain operating experience at intermediate conditions before the final license amendment is approved.

Federal nuclear oversight shapes every phase of a project like Columbia’s uprate, from the NRC’s license amendment authority to the role the Bonneville Power Administration plays as the off-taker whose rate structure absorbs capital costs across its regional customer base.

The second reason is the one that reshaped the budget.

A $500 million scheduling insight, not a scope cut The original EPU estimate was approximately $1.2 billion. The revised figure is approximately $700 million in 2025 dollars. The reduction came from aligning uprate work with lifecycle maintenance that was going to be performed regardless, not from reducing what the project delivers.

That distinction matters for anyone evaluating whether $700 million is good capital allocation. The revised number is structurally lower because the phased approach absorbed roughly half the scope into costs that were already committed.

Reframe it that way and the project stops looking like an expensive standalone uprate. It becomes an incremental cost for a roughly 20% output increase, with much of the work effectively pre-paid through maintenance budgets.

BPA’s planning reflects that staged commitment. Its FY2026 EPU capital allocation sits at $75.578 million, the kind of annual figure that only makes sense when spending is spread deliberately across years rather than concentrated into one cash-hungry shutdown.

Managing the supply chain and workforce complexity that makes or breaks outage schedules

Strategy sets the plan. Procurement and workforce execution decides whether the plan holds, and this is where large nuclear capital projects most often unravel.

The core defence is early procurement. Many of the biggest components carry multiyear manufacturing timelines, so Energy Northwest contracts with original equipment manufacturers well ahead of each outage window to lock in delivery sequencing.

The long-lead categories driving that urgency include:

• Turbine rotors and diaphragms
• Reheaters
• Large motors and pumps
• Generator components

Across the roughly 30 modifications, engineering coordination spans well over two dozen separate technical disciplines, each contributing to the planning and sequencing of work that runs alongside major lifecycle replacement programmes. Raw material commitments for large forgings and castings are placed with suppliers before design is finalised, with contingency sourcing arrangements established where single-supplier exposure would otherwise create schedule vulnerability. Progressive design reviews conducted alongside manufacturers to confirm production feasibility before commitments are made.

One procurement milestone stands out as a signal rather than a routine order.

Learning from Dresden and Quad Cities A contract for steam dryer materials has already been executed. The replacement is driven by experience at other GE boiling water reactors, including Dresden and Quad Cities, where increased steam flow caused high-frequency vibration and fatigue cracking in the original steam dryers.

That early contract tells you Energy Northwest is applying fleet-wide experience rather than waiting to discover the problem at uprated conditions. For anyone tracking execution, pre-emptive steam dryer procurement is one of the clearest indicators that the project team is working from lessons learned rather than theory.

Workforce qualification and execution quality

The staffing model combines internal teams with specialised contractors, and its defining feature is front-loaded engineering. Scheduling uprate work to coincide with biennial refueling outages means skilled-craft demand rises incrementally during those windows rather than requiring a separate, extended shutdown to accommodate the additional scope. Technical specifications are finalised early so contractor teams arrive with resolved designs rather than field-engineering problems on the critical path. That discipline is what keeps an outage from stretching when crews hit the plant.

The constraint to watch is talent availability. Uprate projects need nuclear-qualified welders, turbine specialists, and licensed operators familiar with uprated plant behaviour, and experienced personnel become harder to secure as more plants pursue significant upgrades at the same time.

What 2031 depends on between now and then

Three categories of variable will decide whether this project closes as planned, and the most consequential one sits outside Energy Northwest’s direct control.

That variable is NRC licensing. As of late 2026, the combined MELLLA+/EPU application has not yet been formally submitted. The projected review window runs Q2 2028 through Q4 2030, and any additional NRC requests for information can push implementation back. The clearest single signal of whether the 2031 target holds will be the formal submission and docketing of that combined application.

The NRC licensing review process for extended power uprates sits within a broader federal oversight framework where safety analysis, environmental review, and reactor-specific conditions are evaluated independently before any license amendment is approved.

The primary internal variable is outage scope creep. Because the 30 projects are integrated with planned lifecycle replacements, scope discovered during R-28 or R-29 can extend outage durations, strain project management, and compress the sequencing chain toward its 2031 deadline.

The third factor is regional demand, which raises the stakes on delivery. BPA’s resource planning links the uprate directly to growing loads, explicitly including data centre demand, and the full 186 MWe anticipated increase represents firm baseload capacity that complements variable hydropower. EPU costs are already embedded in BPA’s BP-26 Integrated Program Review forecast assumptions, and Energy Northwest’s $75.578 million FY2026 capital allocation confirms active pre-outage spending.

Here are the signals worth tracking:

• Formal submission and docketing of the combined MELLLA+/EPU application with the NRC
• R-28 outage scope execution in spring 2027
• On-time delivery of long-lead components, particularly the steam dryer
• BPA resource adequacy planning updates tied to regional load growth

The broader industry is watching too. Columbia’s uprate sits alongside other U.S. nuclear projects tracked by the Nuclear Energy Institute, and its execution will serve as a reference point for whether extended BWR uprates remain a viable near-term capacity strategy next to license extensions and next-generation builds. Knowing which variables are still open is the difference between informed monitoring and simply waiting for a press release.

Columbia’s extended power uprate is one of several alternative uprate pathways being pursued across the U.S. fleet; metallic fuel technologies represent a distinct approach that could achieve comparable output gains without the same scale of hardware replacement.

This article is for informational purposes only and should not be considered financial advice. Investors should conduct their own research and consult with financial professionals before making investment decisions. Forward-looking statements regarding project timelines, costs, and capacity targets are speculative and subject to change based on regulatory, market, and operational developments.