# Walsh Promises Faster Nuclear Cleanup. Auditors Found $500m in Contract Growth—and No Reliable End-State Scorecard.

Published: 2026-10-31

Factrail | Investigative accountability analysis | United States | Research cut-off: 31 October 2026

Evidence status: Documentary synthesis of published federal audits, procurement awards, agency responses, contractor descriptions and independent reporting. No allegation of criminal wrongdoing or improper political payments is established.

## The finding: more contract value is not the same as more cleanup

The Energy Department says it is accelerating the cleanup of America's nuclear-weapons legacy. Its own contract records, examined by federal auditors, tell a more qualified story. By March 2025, the department had awarded nearly $20 billion in work under a contracting model built around measurable cleanup outcomes. Only 29 of 57 task orders defined such outcomes. In selected orders, post-award changes increased contract values by more than $500 million; that is not a finding that $500 million was stolen or wasted. At Savannah River, a contractor processed 93% of a specified waste volume while receiving 99% of its target fee. At Hanford, work requirements were changed or carried forward. The accountable question for Energy Secretary [Chris Wright](/people/energy-chris-wright) and cleanup chief Tim Walsh is whether the government can prove that added contract value buys verified risk reduction—not simply a revised promise.

The distinction matters because [the Department of Energy's 20 October anniversary account](https://www.energy.gov/em/articles/one-year-em-marks-historic-progress-under-assistant-secretary-walsh) celebrates real accomplishments under Assistant Secretary Timothy Walsh: uranium tailings moved from Moab, tank waste turned into glass at Hanford, fuel moved into storage at Idaho and infrastructure commissioned at the Waste Isolation Pilot Plant. These achievements should be counted. But an accomplishment list is not a cost-and-outcome reconciliation for the department's contracting system. A [29 July Government Accountability Office audit](https://files.gao.gov/reports/GAO-26-107745/index.html) concluded that the Office of Environmental Management could not yet show whether its End State Contract Model was better than the approach it replaced. That is the narrower, testable accountability finding.

This investigation joins three documentary lines that are often treated separately: the contracting architecture adopted in December 2019; GAO's July examination of 57 task orders and the incentives paid on particular assignments; and the current leadership's October claims of faster, more economical restoration. The comparison does not make Walsh responsible for designing contracts awarded before he entered office. It establishes the choices now available to him and Secretary Chris Wright: publish comparable outcomes, correct pricing and data systems, and demonstrate that a change of political language is accompanied by a change in auditability.

The accountability question also touches Factrail's [fiscal policy sustainability](https://factrail.ai/drivers/fiscal-policy-sustainability) driver: higher authorized contract values require clear reconciliation before a taxpayer can assess fiscal exposure. The [energy security](https://factrail.ai/drivers/energy-security) driver matters because managing legacy radioactive sites safely affects public confidence in reuse of federal land and future energy infrastructure. [Donald Trump](https://factrail.ai/people/donald-trump) leads the administration now responsible for acting on these audit findings, although the contracting system started in 2019 and the sampled performance cannot all be attributed to the current leadership.

## Documentary timeline: a promise made before the current administration

12 December 2019. DOE [awarded the Hanford Central Plateau Cleanup Contract](https://www.energy.gov/em/articles/doe-awards-hanford-central-plateau-cleanup-contract) to Central Plateau Cleanup Company, a joint venture whose parent-company composition has evolved through corporate reorganizations. The government announced a ten-year ordering period and a maximum contract ceiling of $10 billion. The master award could support fixed-price and cost-reimbursement task orders. A ceiling is not cash paid, and the selection of a prime contractor is not a separate competition for every later task.

27 October 2021. DOE [awarded Savannah River Mission Completion](https://www.energy.gov/em/articles/doe-awards-savannah-river-site-integrated-mission-completion-contract) a separate integrated mission completion contract, with an approximately $21 billion ceiling over ten years. DOE said it had received four proposals and selected the winner on best-value criteria. It identified a skilled workforce and the mission of treating and disposing of radioactive liquid waste. Again, the ceiling is not a booked payment or guaranteed corporate revenue.

March 2025. The cutoff used for the July 2026 audit: 57 task orders across nine end-state contracts, with aggregate awarded task-order value of about $19.679 billion. The master contract ceilings in the GAO table summed to about $100.534 billion. The two totals measure different things; combining them or treating the ceilings as obligations would inflate the story. GAO examined the full task-order universe for descriptive purposes, then selected 19 orders for detailed cost review and nine completed orders for outcome review.

2 March 2026. A separate [GAO cost-and-schedule report](https://www.gao.gov/products/gao-26-107820) described a remaining cleanup portfolio exceeding $500 billion and warned that information about major projects and ongoing activities was not consistently comparable. Its reported multiyear increases are not interchangeable with the later $500 million task-order sample. They have different scopes, denominators and baselines.

29 July 2026. GAO published [GAO-26-107745](https://www.gao.gov/products/gao-26-107745). It found that only 29 of the 57 task orders defined end states, that selected completed orders did not consistently meet original outcome terms, and that post-award changes increased values of selected orders by more than $500 million. DOE accepted five of six recommendations and partly accepted the sixth.

26 August 2026. A different [GAO contractor-assurance audit](https://www.gao.gov/products/gao-26-107850) found that field offices sometimes called contractors' self-monitoring systems effective without defined criteria. It documented specific failures in work planning and safety-related reporting. This was a distinct review of three selected field offices, not proof that all contractors or all end-state orders failed.

20 October 2026. DOE marked Walsh's first year as the Senate-confirmed cleanup chief. Its account foregrounded milestones and proposed reuse of former nuclear sites for energy, computing and manufacturing. It did not provide a verified, task-by-task response resolving the July audit's six recommendations. That absence in the published anniversary announcement is not evidence that DOE has done nothing elsewhere.

## How the contract transfers power after the first competition

The contracting model has two decision layers. First, DOE competitively selects a company or joint venture for a large indefinite-delivery, indefinite-quantity master contract. Second, the field office negotiates subsequent task orders with that selected prime. GAO says the model generally does not conduct a new competition at task-order level. A company therefore gains not just an initial award but a seat at the negotiating table for defining later work, pricing and performance incentives. This is an identifiable commercial advantage, not evidence of a rigged award.

The government retains the legal authority to approve scope, choose contract type, set fees, modify work and accept completion. The contractor, however, possesses operational information that can be difficult for the federal customer to replicate. GAO recorded officials explaining that contractors may have more current labor and supplier prices, while federal estimates rely on historical invoices. When the government cannot produce an accurate independent estimate, its formal power to negotiate can be weakened in practice. The crucial relationship is not a politician secretly ordering an award; it is the documented information asymmetry between buyer and seller.

Under a cost-reimbursement order, the government generally carries more cost exposure, subject to allowable-cost rules and negotiated incentives. Under a firm-fixed-price order, the contractor normally bears more performance-cost risk. A cost-plus-incentive-fee arrangement can share gains and losses around a target. GAO observed that selected task orders placing more cost risk on contractors generally had less value growth. That correlation does not prove every fixed-price contract would be appropriate for radioactive cleanup. The technical unknowns can be substantial. It does mean DOE should explain why a particular risk allocation is justified and whether it was revisited when work changed.

The named firms are not interchangeable. [Central Plateau Cleanup Company](https://cpcco.hanford.gov/page.cfm/Overview) identifies an Amentum-led partnership with Fluor and AtkinsRéalis at Hanford. [Savannah River Mission Completion](https://www.savannahrivermissioncompletion.com/about-srmc) identifies BWX Technologies as parent with Amentum and Fluor as partners. Their participation makes them beneficiaries of opportunities to earn fees on federal cleanup work, but the record reviewed here does not establish their actual net profits, distributions to shareholders or the amount of any questionable payment. The GAO findings also do not assign all of the portfolio's $500 million increase to either of those two companies.

## The Hanford test: 22 outcomes, twelve rewritten scopes

The most instructive case is not the largest percentage increase. It is what counts as an outcome when the target itself moves. At Hanford, GAO reviewed a task order valued at about $1.8 billion and identified 26 distinct cleanup end states. The documentation supported completion of 22. Yet DOE had changed the scope for twelve of those 22, one end state was not achieved, and the auditors could not establish whether the remaining three were complete. Some work was carried into a later task order. The number 22 therefore cannot be presented as 22 unchanged promises delivered exactly as sold.

This is not necessarily a scandal. A contaminated site is not a standard road project. Unexpected radioactive conditions, changing safety requirements and dependencies between groundwater treatment, demolition and waste disposal can make a revised scope sensible. GAO records Hanford officials explaining that interdependent work caused them to combine activities originally distributed across separate orders. But those practical reasons make an auditable change log more important, not less. A taxpayer should be able to see the original physical result, the revised physical result, the money added or released, the schedule change and the residual liability left for the next order.

The [Engineering News-Record account of the July audit](https://www.enr.com/articles/63394-gao-finds-cost-growth-weak-oversight-under-doe-nuclear-cleanup-contract-model) highlighted a further consequence: DOE transferred work worth roughly $19 million into a later order. That is a change in the assignment of work, not proof that $19 million vanished. The public-interest issue is whether the same work is counted as progress in one order while its cost and completion risk migrate into another. GAO said the records did not clearly allocate fees to individual Hanford end states. Without that bridge, it is difficult to determine what was paid for each original promise.

Hanford's community and state regulators have an additional stake. [Local reporting in April](https://www.tricitiesbusinessnews.com/articles/hanford-unsustainable-budgets) described a record fiscal 2026 federal allocation alongside Washington state's concern that funding remained short of what would be needed to satisfy legal cleanup milestones. A large appropriation can be welcome and still insufficient. Neither a contractor's claimed completion percentage nor a Washington budget demand alone settles which remediation activities are safe, lawful and achievable.

## The Savannah River test: 93% of the volume, 99% of the target fee

At Savannah River, the fee mechanism is unusually revealing. One selected task order required processing 4.5 million gallons of liquid waste. GAO found that approximately 4.2 million gallons were processed, about 93% of the specified volume. The contractor received a $67 million fee—99% of the target fee, though only 78% of the total available performance-based incentive fee. Reporting the 99% figure without its denominator would exaggerate the case, and reporting the 78% figure alone would conceal the shortfall against the original volume target.

DOE's explanation is substantive. The contract deliberately weighted incentives toward earlier, technically harder and higher-risk stages of processing. According to GAO, the fee design offered a 7% fee for the first 2.5 million gallons and a 7.5% fee for the remaining 2 million. Federal officials said this was consistent with paying for complexity and risk rather than only for the final gallon. That is the strongest available counterargument to a simplistic assertion that the contractor was paid for nothing.

But the structure creates a public accountability question. If most of the reward is available before the final outcome, who bears the cost and delay when the final portion is deferred? The answer cannot be established from the percentage alone. It requires the task-order amendments, subsequent waste-processing records, actual payments and any adjustments to later fees. GAO's case demonstrates that a performance fee can be lawful under its terms while still leaving an important portion of the original environmental outcome incomplete.

The comparison with two other selected orders matters. GAO found a Savannah River order for two saltstone disposal units that met its defined approvals, one early and the other slightly late relative to a task deadline but early against the overall project schedule; it earned its full $4.8 million fee. An Oak Ridge site-preparation order also finished as planned and earned its full approximately $374,000 fee. The record therefore contains successes as well as failures. The finding is inconsistency and inadequate cross-site measurement, not universal contractor nonperformance.

## The $500 million: what grew, what was paid, what cannot be inferred

GAO's headline amount measures increases in the value of selected task orders between initial award and the March 2025 comparison date. It does not mean the government had already paid every added dollar, that all growth represented profit, or that every change was avoidable. A more expensive contract can reflect additional legitimate cleanup work. A contract can also become more expensive because an original estimate was poor, negotiations started before price was fixed, or the government transferred cost risk to itself.

Specific examples expose the spread. The audit listed roughly $216.7 million of growth on an Oak Ridge phase-in order, about $75.8 million on a Nevada environmental-operations order, $75.6 million on an Idaho hybrid integration order, and $25.8 million on Oak Ridge mercury-treatment work. The percentages ranged widely. Those figures are not a league table of misconduct. They identify where procurement officials should reconcile new scope, estimated cost, negotiated fee and actual outcome.

Six selected task orders began as undefinitized contract actions: work started before scope, cost, schedule and fee had all been finalized. GAO said four exceeded the generally applicable 180-day definitization deadline. In that interval the government's leverage can weaken because costs are being incurred before the full price is settled. DOE may have valid operational reasons to start urgent work, but the mechanism shifts uncertainty toward the public purse.

A second audit from August adds a different warning. [GAO-26-107850](https://www.gao.gov/products/gao-26-107850) found that Hanford and Idaho field offices rated contractor assurance systems effective without defined criteria, while Los Alamos had not made a determination. The review cited improperly closed or inadequately documented significant issues at Hanford, reported radiological events at Idaho, and a training-related stop-work episode at Los Alamos. These are not the same contracts or a single $500 million finding. They show why confidence in a contractor's internal reporting cannot replace a federal test with clear definitions.

## Who can act, who can block, and who benefits

Congress appropriates money and can condition it through statutes and oversight. Its committees requested the GAO work. Lawmakers can press for data and demand hearings, but an individual senator does not personally approve a Hanford task-order fee. The secretary of energy, Chris Wright, has department-wide political and managerial responsibility. Assistant Secretary Walsh leads the Office of Environmental Management and is the official to whom GAO directed its recommendations. Field contracting officials negotiate and administer the particular task orders. GAO audits and recommends; it cannot rewrite an executed contract. State regulators and courts can enforce environmental obligations through applicable legal instruments. Contractors execute the work, submit estimates and earn allowable fees under the negotiated terms.

The direct commercial beneficiaries of awarded contracts are the selected prime contractors and their subcontracting networks. A master contract gives access to work and potential revenue. The public can benefit from safer water, stabilized waste, local jobs and usable land when milestones are genuinely delivered. The public can also carry additional fiscal risk when prices are uncertain and deliverables are revised. Neither benefit should be converted into a fictitious corporate windfall or a fictitious dollar loss without payment and cost data.

The October DOE announcement also promotes announced private investment in redevelopment of former nuclear sites. These are distinct projects with different investors and approval paths. An announcement of private capital is not evidence that cleanup contractors earned it, that it has all been spent, or that it offsets federal environmental liabilities dollar for dollar. Mixing those totals with nuclear-cleanup obligations would conceal rather than clarify who pays.

## The strongest counter-case: nuclear remediation is not a fixed-price commodity

DOE's formal reply to the July audit disputes the premise that every increase in contract value should be treated as a cost-control failure. The department noted that federal procurement regulations do not define all such changes as cost growth and that new requirements can reflect the complexity of contaminated sites. It agreed with five recommendations and partly agreed with the recommendation to identify root causes of growth. Its proposed approach would focus analysis on orders with unsatisfactory cost-control performance in the contractor rating system.

GAO's response is persuasive but narrower than an accusation of corruption. Contractor ratings can measure contractor behavior; they do not necessarily reveal whether the government itself wrote inadequate requirements, prepared a weak independent estimate or made a late decision that increased cost. GAO therefore wants root-cause review of all growth, not only cases already flagged by a contractor-performance score. That is a dispute about how to govern public risk, not a finding that DOE officials knowingly overpaid.

Walsh can also point to genuine progress. DOE's October account reports completion of Moab tailings removal, a first Hanford glass-production milestone, fuel transfers and facility improvements. The [American Nuclear Society's reporting on March's GAO assessment](https://www.ans.org/news/2026-03-05/article-7820/doe-nuclear-cleanup-costs-schedule-delays-continue-to-rise-gao-says/) recognizes both the extraordinary long-term scale and opportunities for savings. The [Hanford grouting assessment](https://www.gao.gov/products/gao-26-108878) shows why apparently cheaper treatment options require scrutiny of transport, disposal, permits and waste characteristics. A serious audit should not punish an agency merely for discovering that a dangerous cleanup is more complex than first estimated.

## What the records would have to show next

The first test is a public, versioned record for every material task-order change: original end state, modified end state, why it changed, which official approved it, revised cost ceiling, payments to date, earned fee and remaining environmental risk. The second is a reconciliation between the two federal procurement systems GAO found inconsistent, with subtask-level changes made visible where technically possible. The third is a headquarters dashboard that distinguishes completed physical outcomes from planning work and does not treat a moved milestone as an unchanged success.

The fourth test is the department's response to GAO's open recommendations. The July [recommendations tracker](https://www.gao.gov/products/gao-26-107745) lists six as open, including more consistent reporting, oversight of outcomes, risk-reduction planning, cost-growth root causes, independent estimates and timely definitization. DOE has promised corrective steps; a promise is not an independently verified closure. The fifth test is whether Walsh's redevelopment strategy discloses the separate federal cleanup obligations that remain after private investors announce projects.

There is no documented basis here to accuse Walsh, Wright or the named contractors of bribery, fraud or a secret exchange. Much of the audited performance predates the present leadership. There is a documented basis to demand a verifiable answer to a different question: when the government celebrates an end state, can citizens identify the original promise, the actual work completed, the cost added, the fee earned and the radioactive risk that remains?

## Sources and evidence boundaries

The core audit is [GAO-26-107745, July 29, 2026](https://files.gao.gov/reports/GAO-26-107745/index.html), with the [official recommendation status](https://www.gao.gov/products/gao-26-107745). Related primary checks are [GAO's March cost-and-schedule report](https://www.gao.gov/products/gao-26-107820), [August contractor-assurance report](https://www.gao.gov/products/gao-26-107850), [Hanford grouting report](https://www.gao.gov/products/gao-26-108878), [DOE's 2019 Hanford award](https://www.energy.gov/em/articles/doe-awards-hanford-central-plateau-cleanup-contract), [2021 Savannah River award](https://www.energy.gov/em/articles/doe-awards-savannah-river-site-integrated-mission-completion-contract), and [20 October 2026 leadership statement](https://www.energy.gov/em/articles/one-year-em-marks-historic-progress-under-assistant-secretary-walsh). Independent reporting consulted includes [Engineering News-Record](https://www.enr.com/articles/63394-gao-finds-cost-growth-weak-oversight-under-doe-nuclear-cleanup-contract-model), [American Nuclear Society](https://www.ans.org/news/2026-03-05/article-7820/doe-nuclear-cleanup-costs-schedule-delays-continue-to-rise-gao-says/), and [Tri-Cities Area Journal of Business](https://www.tricitiesbusinessnews.com/articles/hanford-unsustainable-budgets). The [Hanford](https://cpcco.hanford.gov/page.cfm/Overview) and [Savannah River](https://www.savannahrivermissioncompletion.com/about-srmc) contractor sites establish organizational roles, not independent proof of performance.

Unresolved: Complete contract modifications, actual cash disbursements and fee adjustments, subcontractor ownership and margins, independent remediation outcome series, subsequent progress on GAO recommendations, and verified Factrail canonical links. No publisher or official named above has supplied a new on-record response specifically to this article; existing published replies are described on their own terms. Further independent legal review of any new misconduct claim would be required; this report makes no such claim.

## Methodology

https://factrail.ai/methodology

Canonical: https://factrail.ai/articles/doe-nuclear-cleanup-contract-growth-end-state-walsh-2026