US Dam Retrofits Could Add 15.2 TWh: What the Data Shows

Nearly all of America’s dams generate no electricity. Roughly 97% of the country’s dam infrastructure sits without a single turbine, moving water for flood control, navigation, or water supply while producing nothing for the grid. A new federal laboratory study just put a precise number on what that gap is worth.

Published on 27 July 2026 by Oak Ridge National Laboratory (ORNL) in collaboration with Idaho National Laboratory (INL), the assessment is the most granular federal read yet on what retrofit technology and improved hydrologic modelling reveal about existing dam infrastructure. A software platform called HydroGenerate moves the analysis from a broad observation about untapped potential to a site-by-site calculation.

This matters now for three reasons: clean capacity additions are in demand, grid resilience is a policy priority, and federal infrastructure funding has specifically targeted this pathway.

Here is what the data actually tells you: the difference between technical potential and a realistic project pipeline, where the geographic and policy levers sit, and how to read the headline 15.2 TWh figure with the precision it deserves.

What the new federal assessment actually found

The headline is genuinely striking. A subset of more than 2,600 non-powered dams across the United States could add roughly 4 GW of capacity and generate 15.2 TWh of electricity annually, enough to power more than 1.4 million homes, all without building a single new dam.

15.2 TWh a year from existing dams Enough electricity to power more than 1.4 million American homes, achieved entirely by retrofitting infrastructure that already exists.

The average candidate facility comes in at around 1.5 MW, which tells you this is a distributed opportunity built from many modest projects rather than a handful of giants. Precision is what separates this assessment from the national-scale approximations that preceded it, and the tool doing the precision work is itself part of the story.

HydroGenerate is a multi-stream modelling platform developed at Idaho National Laboratory under principal investigator Carly Hansen, who led the report. Rather than applying broad national averages, it produces a site-level estimate for each dam by combining several distinct data streams.

ORNL’s hydropower assessment provides the site-level methodology behind HydroGenerate, including how the model integrates USGS flow data, elevation datasets, turbine performance specifications, and hydraulic head measurements to produce facility-specific generation estimates rather than national averages.

The model integrates four core inputs:

• U.S. Geological Survey (USGS) water flow data
• ORNL elevation datasets and infrastructure inventories
• Turbine performance specifications
• Hydraulic head measurements (the vertical distance water falls, which drives how much power a turbine can produce)

It also accounts for seasonal flow variation and the structural realities of each site, including flood-control and navigation obligations that limit how much water can actually be routed through a turbine.

Here is the calibration point you need. The 15.2 TWh figure is a modelled technical potential that assumes full development of every identified site, not a deployment forecast. ORNL frames it as an estimate-driven upper bound, which means it describes the ceiling of what is physically possible rather than what economics and regulation will allow to be built. Everything that follows in this analysis is about the distance between that ceiling and the ground.

Where the opportunity is concentrated, and why the map looks the way it does

The intuitive assumption is that America’s hydropower future lies where its hydropower present already sits: the Pacific Northwest. The data points somewhere else entirely.

U.S. hydropower is heavily clustered in a few western and northwestern states. According to 2025 figures from the U.S. Energy Information Administration (EIA), the top five states account for 60% of the country’s conventional hydroelectricity net summer generation capacity.

The retrofit candidates tell a different geographic story. They concentrate in the Great Lakes watershed and the upper Mississippi River region, alongside major navigation dams on the Ohio and Tombigbee systems.

The reason is structural, not coincidental. The Pacific Northwest’s major federal dams on the Columbia and Snake rivers were fitted with generators decades ago, so there are simply few large non-powered dams left there to convert.

The Midwest and South are different. They hold large Army Corps of Engineers navigation and flood-control dams that were built to move barges and manage water, never to make electricity, and these represent the largest untapped flow resources in the country. Just 10 large non-powered dams on the Ohio and Tombigbee systems could provide approximately 3,000 MW. Widen the frame to the top 100 sites and the figure reaches roughly 8,000 MW, still skewed toward central and eastern river systems.

What this tells you is that the retrofit opportunity would bring clean capacity to regions that currently have almost no access to hydropower, a diversification of the grid rather than a reinforcement of where it is already strong.

Leveraging hydropower for electricity access is a policy argument that appears in both developing-world energy strategy and U.S. regional grid planning, and the geographic equity dimension of Midwest dam retrofits, bringing clean capacity to regions with minimal existing hydropower, closely mirrors the access rationale driving large-scale hydro investment in other contexts.

The federal ownership factor

One number reframes the entire opportunity: 86% of retrofit candidates are federally owned facilities.

That concentration of ownership means progress on these projects is fundamentally a federal policy and permitting question, not primarily a private capital deployment question. Decision-making authority sits with Army Corps of Engineers permitting, Federal Energy Regulatory Commission (FERC) licensing, and Department of Energy (DOE) funding programmes rather than with independent developers.

The practical implication for you is twofold. Federal coordination requirements add complexity and lengthen timelines, but they also create a pathway where targeted federal investment can unlock projects that market economics alone would never support. Changes in federal priorities therefore have outsized consequence for how much of the modelled potential ever gets built.

The real cost of a retrofit, and why most sites stay unpowered

The cost data is where technical potential meets financial reality, and it is best read not as a discouraging obstacle but as the lens that makes the gap between 15.2 TWh and a realistic pipeline legible.

The most current national benchmark comes from the National Renewable Energy Laboratory (NREL) 2024 Annual Technology Baseline, which puts the overnight capital cost of non-powered dam retrofits at $3,045 to $20,043 per kilowatt. ORNL’s earlier empirical work, drawn from nearly 20 real sites, stretches the range even wider.

That dispersion is the point. When real sites range from under $2,000/kW to more than $52,000/kW, the viability of any specific retrofit is almost entirely site-dependent, and no headline potential figure can be evaluated without site-specific cost analysis behind it.

Three factors drive the cost:

Hydropower turbine technology has advanced considerably since most U.S. navigation dams were built, and modern electromechanical equipment from suppliers serving large-scale projects represents a materially different cost and performance profile than the baseline assumptions embedded in older site assessments.

• Water conveyance structures (the channels and penstocks that route water to the turbines)
• Powerhouse construction
• Electromechanical equipment (turbines, generators, and control systems)

Lake dams and navigation dams tend to carry particularly high conveyance costs, which is why some of the geographically attractive Midwest sites also sit at the expensive end of the curve. A 7.6-MW test facility retrofit, for context, was estimated at roughly $58 million in 2021 dollars, illustrating how a single-digit-megawatt project can reach tens of millions.

So why pursue retrofits at all when the costs run high? Because relative to the alternative, they are efficient.

Retrofits can cost 20-60% less than new-build hydropower DOE and ORNL analyses suggest converting an existing dam avoids much of the capital burden of greenfield construction, even where absolute costs remain elevated.

Set against utility-scale solar or land-based wind at current NREL benchmarks, non-powered dam retrofits are still substantially more expensive per kilowatt. What that means for you as an investor is direct: most sites will not clear financial hurdles on market economics alone, which is precisely why the federal incentive structure exists.

The clean capacity investment landscape in 2026 is shaped by competing priorities across solar, wind, geothermal, and hydropower, and dam retrofits occupy a distinct position within that mix: higher per-kilowatt costs than solar or wind, but with dispatchability advantages and a geographic diversification argument that neither of those technologies can replicate.

Federal funding, FERC complexity, and the gap between policy intent and project reality

The policy support for these retrofits is real and specific. The frictions that keep sites from progressing are equally real, and the honest read sits in the tension between the two.

Federal backing is anchored in the Infrastructure Investment and Jobs Act (IIJA), also known as the Bipartisan Infrastructure Law. The National Hydropower Association characterised the combined waterpower commitment as a $909 million “down payment.”

Beyond those pools, DOE’s Water Power Technologies Office released a dedicated funding opportunity, DE-FOA-0002731, announced on 10 May 2023, making $14.5 million available with a topic area targeting non-powered dam retrofits directly. On 6 September 2023, DOE announced more than $13 million for seven hydropower research and development projects under the Bipartisan Infrastructure Law.

Read that funding architecture carefully and it reveals its own diagnosis. The FOA explicitly funds “studies and designs that facilitate the licensing, construction, and commissioning” of retrofits, which is a tacit admission that FERC licensing is the central bottleneck.

The licensing process is resource-intensive and lengthy, and environmental review adds further delay. When the federal government allocates dedicated money simply to get projects licensing-ready, it is telling you that the paperwork gap is as material as the funding gap.

Energy permitting reform is the single variable most capable of closing the gap between modelled potential and operating capacity, since FERC licensing timelines govern project delivery more directly than capital costs or hydrologic suitability at any individual site.

The structural barriers cluster into four:

1. Aging infrastructure and site-specific engineering complexity, since most dams were never designed to generate power
2. High and variable capital costs, with per-kilowatt figures swinging by more than an order of magnitude
3. FERC licensing and environmental review, the central permitting gateway
4. Financing gaps and reliance on incentives, because unsubsidised retrofits struggle to attract private capital

What this means for you is that production incentives and licensing-readiness funding are not generosity; they are a structural acknowledgement that unsubsidised retrofits cannot clear financial hurdles at most sites. There is one further caveat. IIJA programmes were active through 2023-2024, but the broader federal clean energy investment direction was shifting as of September 2026, so current support should be treated as a snapshot rather than a guaranteed ongoing commitment.

Calibrating the 15.2 TWh figure for a clean energy investor

Stack the analytical layers together and a calibrated read emerges, one that is neither dismissive nor credulous.

The 15.2 TWh across 4 GW and more than 2,600 sites is a meaningful ceiling, not a pipeline. The realistic near-term opportunity is far narrower, concentrated in a subset of large, federally owned navigation dams across the Midwest and South.

The headline is heavily top-weighted Just 10 non-powered dams on the Ohio and Tombigbee systems represent approximately 3,000 MW, meaning a small number of large-site decisions could drive a disproportionate share of any realistic build-out.

That concentration is the key strategic insight. Because 86% of candidates are federally owned and the potential is so top-heavy, policy decisions affecting a handful of Army Corps navigation dams could determine whether this pathway delivers at scale or stays an analytic curiosity.

Three variables will most directly shape how much modelled capacity becomes operating capacity:

1. FERC licensing and federal permitting reform, which governs project timelines more than any other factor
2. Continuation of Section 242 and 243 incentives beyond current appropriations
3. Site-specific cost outcomes on the first wave of large-dam retrofits, which will set the template for what follows

There is also a durable non-economic argument worth weighting. Retrofits would bring clean capacity to the Great Lakes and upper Mississippi regions, areas with minimal existing hydropower, and the retrofit cost advantage of 20-60% below new-build remains intact even where absolute costs are high. That geographic equity dimension gives the pathway a bipartisan rationale distinct from the pure economics, which matters when federal priorities are the deciding lever.

What the data changes, and what it does not

The 27 July 2026 ORNL-INL assessment genuinely advances the picture. It is the most precise site-level national analysis to date, it narrows the uncertainty around the modelled potential, and it hands FERC, the Army Corps, and private developers the kind of inventory-level specificity they need to prioritise which sites are worth pursuing first.

What it does not do is change any of the conditions that keep those sites unpowered. It does not shorten the FERC licensing timeline. It does not lower site-specific capital costs. It does not guarantee federal funding continuity beyond current appropriations.

The takeaway is a clean one. Better modelling tools lower information risk, but they do not lower execution risk, and the gap between what HydroGenerate can calculate and what actually gets built will be closed by policy and financing decisions rather than by further analysis.

The 97% figure remains the anchor. The study clarifies exactly what the untapped opportunity is worth, but the structural conditions leaving nearly all U.S. dams unpowered are still firmly in place. The enduring case for retrofitting large federal navigation dams in the Midwest and Great Lakes rests as much on grid diversification and regional energy access as on the numbers alone.

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. Past performance does not guarantee future results, and financial projections are subject to market conditions and various risk factors. Forward-looking statements regarding technical potential and federal funding are speculative and subject to change based on policy and market developments.