Wright Oversees a Nuclear Cleanup Estimated at Up to $840bn. Auditors Found “Effective” Safety Ratings Without a Test.

Factrail investigation | United States | Documentary trigger: 26 August 2026 | Researched: 1 November 2026

Search question: How did the Department of Energy rate nuclear-cleanup contractors' safety assurance systems as effective without measurable criteria, and who can change that?

The finding

The United States is paying private companies to remove radioactive and chemical legacies of its nuclear weapons programme, while relying partly on those same companies to identify and report failures in their own work. In an August 2026 Government Accountability Office investigation (opens in a new tab), two federal field offices had described contractors' assurance systems as effective, yet auditors could not establish the criteria supporting those judgments. At three sampled sites, GAO found no clear contractual consequences specifically tied to poor performance of the assurance systems. One contractor had documented 15 radiological-control events in eight months; another had closed or inadequately documented 13 of 33 sampled high-significance issues; a third suffered a training breakdown that delayed cleanup for 90 days. These are documented oversight weaknesses, not proof that all three contractors are unsafe, that any politician received money, or that the government's projected cleanup bill was lost. The accountability question is whether Energy Secretary Chris Wright and Environmental Management chief Timothy Walsh will turn information audit findings into enforceable, measurable oversight.

What is new in this examination

The distinctive finding is a comparison of three records that are usually discussed separately: the August 26 assurance-system audit (opens in a new tab), a May 19 staffing audit (opens in a new tab), and the department's own subsequent accounts of cleanup progress. Together they expose a difficult policy choice. Federal supervisors were relying on contractor-generated information at the same time that the oversight organisation was severely understaffed, and the contracts reviewed did not make the quality of that self-reporting a clearly measurable financial condition. The evidence does not show that staffing losses caused any particular safety event; most of the specific failures described by GAO predate the 2025 workforce contraction. It does show that a future strategy relying even more heavily on contractor self-assurance needs a credible independent test.

The GAO's August report (opens in a new tab) puts the entire Environmental Management mission at 15 contaminated sites, with an April 2026 estimate of $641 billion to $840 billion of remaining cleanup work potentially extending to 2100. This is a long-range estimate of future programme costs, not an annual appropriation, current contract award or amount reported to be misspent. The same audit says approximately 95% of the office's roughly $8 billion annual budget goes to contracting, supporting a contractor workforce exceeding 25,000. The economic exposure is therefore both unusually large and unusually long-lived.

The documentary timeline: the warning preceded the latest audit

The underlying federal obligation did not begin with the current administration. DOE Order 226.1B requires federal oversight and contractor assurance systems, or CASs. These are supposed to combine credible self-assessment, issue classification, reporting to contracting officers, feedback, corrective action and performance measures. A well-designed system can make workers safer by detecting weak practices before an accident. It is not inherently improper for a contractor to assess its own operations; the failure arises if federal officials cannot independently determine whether that assessment is trustworthy.

In 2022, the Defense Nuclear Facilities Safety Board questioned whether DOE had sufficiently documented the reliability and effectiveness of the contractor-assurance information on which its oversight depended. The August 2026 GAO report (opens in a new tab) records that warning and DOE's agreement that clearer evaluation criteria should be developed. GAO says Environmental Management headquarters officials were unaware in January 2026 of concrete action to create such criteria. That is a chronology of an unresolved management weakness across administrations, not an allegation that Wright personally created it.

On 24 April 2024, DOE's independent Office of Enterprise Assessments (opens in a new tab) published an assessment of nearly 4,000 safety issues managed by nine nuclear-facility contractors over fiscal years 2019–2023. GAO's later account says the assessment found about a quarter of sampled issues inadequately managed. It also describes four contractors misclassifying hundreds of nuclear-safety deficiencies in ways that could avoid or delay required action. The 2024 assessment is a separate internal independent-review lineage, not nine additional cases that GAO personally sampled in 2026.

The August 2026 audit itself drew on documents, interviews, an Idaho site visit and a non-representative sample of three contracts: Idaho Environmental Coalition at Idaho, Newport News Nuclear BWXT-Los Alamos at Los Alamos, and Central Plateau Cleanup Company at Hanford. It reviewed performance evaluations from fiscal years 2022–2024 and older effectiveness declarations, among other evidence. Its conclusions cannot be converted into a national percentage of failed contractors.

The 2025 political transition added a new managerial responsibility. DOE identifies Chris Wright as energy secretary (opens in a new tab), confirmed on 3 February 2025; the Associated Press reported (opens in a new tab) his confirmation and energy-policy priorities. DOE says Timothy Walsh became assistant secretary for Environmental Management in October 2025 (opens in a new tab). Both inherited earlier oversight arrangements. Neither can reasonably be blamed personally for the 2023 training failure solely because they now occupy senior offices. They can be held accountable for what happens to the August 2026 recommendations under their authority.

Idaho: fifteen events, and an ambiguous definition of success

At Idaho, the field office's effectiveness declaration said its oversight processes implemented the CAS requirements and that work was being performed safely. But the GAO's detailed findings (opens in a new tab) also describe an earlier hazardous-energy-control violation and a recurring pattern of 15 radiological events between June 2024 and February 2025. The contractor, Idaho Environmental Coalition, documented events that put personnel at risk of radiological overexposure and could have spread contamination. Examples included missing dosimeters or failures to respect radiological boundaries.

The distinction between risk and injury is essential: GAO did not say that 15 workers received excessive doses or that contamination actually spread in each incident. The agency's own discussion makes the evaluative problem explicit. A system that identifies a pattern might be working; a system that allows fifteen recurrences before recognising the pattern might be inadequate. Without an agreed threshold, the same evidence can support contradictory ratings.

Idaho Environmental Coalition's own description (opens in a new tab) identifies an Amentum-led team with North Wind Portage and specialist subcontractors. A 2021 Jacobs award announcement (opens in a new tab) originally described Jacobs as the majority partner, illustrating why ownership and contractor branding must be dated rather than assumed static. The 2021 award described a ten-year $6.4 billion estimated contract value. That ceiling or estimate is not a payment information, profit calculation or proof of an improper award. The contractor also points to completed cleanup milestones and protection of the Snake River Plain Aquifer. Those outcomes are relevant to its strongest defence, but milestone performance is not a substitute for validating the quality of radiological-risk reporting.

Hanford: “effective” alongside poorly closed findings

At Hanford, the Central Plateau Cleanup Company and field office likewise characterised the assurance system as effective or fully implemented. The same materials recorded poor work planning and conduct of operations, with federal requests for corrective action. A January 2024 contractor review found that 13 of 33 sampled issues in the two highest significance categories, drawn from January 2022 to October 2023, had been improperly closed or inadequately documented. This is roughly 39% of that selected sample, not 39% of all Hanford issues, all nuclear hazards or all contractors.

The GAO report (opens in a new tab) adds a revealing denominator problem. Central Plateau reported 2,734 issues in fiscal 2024 and 1,908 in fiscal 2025. The Hanford field office tracked 346 and 320 respectively. Environmental Management headquarters tracked just 24 Hanford issues across 2024 and 2025 in its own selected issues-management system. These are different systems and selection criteria, so subtracting the figures would not reveal a number of concealed incidents. What matters is that headquarters could not demonstrate a documented, risk-based selection method for deciding which signals should shape its oversight.

The contractor's ownership and financial incentives are identifiable. DOE's 2019 award notice (opens in a new tab) named the Central Plateau Cleanup Company consortium and a $10 billion maximum contract ceiling over a ten-year ordering period. The contractor's current site (opens in a new tab) identifies Amentum, Fluor and Atkins as its experienced partnership. A maximum ordering ceiling is neither a disbursement nor the partners' realised earnings. Those companies benefit commercially from having the right to perform federal cleanup work; their exact fees, margins and payments attributable to assurance-system ratings require contract-level accounting not established here.

The counter-record is substantial. On 13 October 2026 DOE reported (opens in a new tab) that Central Plateau crews had stabilised contaminated gloveboxes at Hanford's Waste Receiving and Processing facility, projecting approximately $8 million in annual surveillance and maintenance savings. This is an agency forecast, not an independently reconciled saving. Nevertheless, it describes real risk-reduction work that a fair account must not erase. Successful field projects and a weak method for assessing contractor self-oversight can coexist.

Los Alamos: a ninety-day delay, then evidence of progress

The third case is Newport News Nuclear BWXT-Los Alamos, or N3B. Its field office did not declare the CAS effective; it reported concerns. A 2023 review commissioned by N3B management found that the system was less mature than expected, including reluctance to register issues and pressure to classify them as less significant. GAO describes a separate 2023 training and qualification breakdown affecting three nuclear facilities. It produced a stop-work order and a 90-day delay to nuclear-waste disposal and environmental remediation. That is a documented operational consequence, but the report does not quantify its net dollar cost or prove deliberate suppression by named executives.

The financial oversight question is especially sharp here. GAO found that performance documents for fiscal 2022–2024 did not isolate the portion of subjective fees linked to CAS performance. Contracting officers did not recall taking a contractual action or downgrading an evaluation specifically because of CAS deficiencies. Indeed, GAO found a fiscal 2023 fee narrative praising improved participation in assurance activities even as the contractor-commissioned review that year criticised system maturity. These records are not proof that no fees were ever withheld for any reason; they are evidence that GAO could not attribute a fee consequence to the assurance failures.

Nor is a 2023 stop-work episode an accurate description of every later year. In December 2025 DOE reported (opens in a new tab) that N3B and the field office met all eleven fiscal-year cleanup milestones and shipped more than 190 cubic metres of transuranic waste to permanent disposal. The contractor said its injury performance compared favourably with a high-hazard industry benchmark. Those claims warrant checking against raw injury and shipment records, but they are relevant disconfirming evidence. The conclusion is not that Los Alamos cannot deliver cleanup; it is that delivery milestones and CAS quality need separate, independently auditable tests.

Who has the power to change the incentives?

The formal chain runs from Congress, which authorises and appropriates funding and can demand testimony, to the energy secretary, to Environmental Management headquarters, to field offices and federal contracting officers. Contractors operate facilities and maintain assurance systems. They do not hold the legal power to approve their own contract amendments, set the government's award-fee terms or decide the federal oversight budget. The secretary and assistant secretary can direct management improvements within existing law and resources; contract changes must still follow applicable procurement rules and existing contractual rights.

The August GAO report (opens in a new tab) identifies four specific management decisions available to Environmental Management: define effectiveness using measurable criteria; specify and standardise the information needed from sites; actually implement risk-informed headquarters assessment schedules; and establish clear CAS expectations and consequences in contract mechanisms. DOE concurred with all four recommendations and said it planned processes, standardised metrics, field-office assessments and a review of contract mechanisms. Concurrence is not evidence that the actions have already been completed.

The strongest institutional veto is practical as well as legal. A contracting officer may be constrained by the existing agreement, performance-evaluation rules and evidence available to justify withholding a fee. Federal supervisors also face staffing limits. The May 2026 GAO workforce report (opens in a new tab) found that Environmental Management staffing fell from 1,272 in fiscal 2023 to 856 in fiscal 2025. Its reported 45% vacancy rate used the office's 2023 estimate of 1,515 needed positions as the denominator. Of 409 departures in fiscal 2025, 312 occurred through the Deferred Resignation Program. Those numbers do not prove that every vacancy was caused by the current president's policies, nor do they tell us how many vacancies remained in November 2026. They do show the scale of the documented capacity problem during the period the new leadership inherited.

The Tri-City Herald's May 2026 reporting, republished by the Spokesman-Review (opens in a new tab), described 96 federal staff departures in the Tri-Cities Hanford oversight workforce, a 37% reduction, citing the workforce audit. The page was visible in search results but returned an access restriction during direct retrieval; the national GAO figures are independently accessible. A claim that any particular accident resulted from these departures would need a site-level investigation, not a staffing correlation.

The money map: access, fees and who bears risk

There are at least four separate money flows. Congress funds DOE; DOE obligates money under cleanup contracts; contractors receive reimbursements and may earn performance-based fees; and the public ultimately finances the cleanup and bears the consequences of inadequate remediation. These flows should not be collapsed into a claim that the combined $18.5 billion estimated value or ceiling of the three selected contracts is money already spent. GAO listed approximately $6.4 billion for Idaho, $2.1 billion for Los Alamos and $10 billion for Hanford as of November 2024. Their sum is a comparison of contract scales, not a single-year expenditure or profit pool.

The commercial advantage of a contract is concrete: selected firms have access to federal work and potential fees. The narrower possible advantage of a weak assurance metric is more difficult to quantify. If an assurance failure is not separately assessed, a contractor may face less direct fee exposure than under an enforceable metric. But that is a risk of incentive design, not evidence that Amentum, Fluor, Atkins, N3B or another named company actually obtained a specific unearned payment. To prove a financial windfall, a reporter would need the relevant fee plans, actual award decisions, deductions, costs and any counterfactual assessment.

Workers have a different exposure: radiological and industrial safety. Residents near Idaho's aquifer, Hanford's Columbia River corridor and Los Alamos' contaminated groundwater have an interest in competent cleanup and credible monitoring. Their losses are not established by multiplying the number of reported safety events by an invented monetary damage figure. Delay, potential contamination and public mistrust are real categories of consequence, but the size of each must be demonstrated case by case.

For elected officials, the relevant financial tie is public contracting and control of appropriations, not an reported private interest. No inspected source in this record demonstrates a political donation, personal shareholding, lobbying exchange or corrupt instruction connecting Wright, Walsh or congressional overseers to a specific contractor award. The record supports institutional responsibility and named corporate incentives; it does not support a corruption allegation.

The strongest case against the critical interpretation

The contractors and DOE have arguments that cannot be dismissed. Nuclear remediation is technically hazardous, and a self-reporting system may be effective precisely because it identifies and escalates problems rather than pretending they do not exist. More reported incidents may sometimes reflect better detection. Hanford's field office has already developed sitewide standardised terms, which GAO considered useful. The contractor-led Energy Facility Contractors Group survey cited by GAO found that both federal and contractor respondents wanted clearer standards, suggesting a shared technical problem rather than a one-sided refusal to comply.

GAO also reports officials' concern that crude numerical incentives can backfire. Rewarding fewer recorded issues could encourage underreporting; rewarding more assessments could favour paperwork volume over quality. A sensible contract measure would test timely escalation, accuracy of classifications, independently validated corrective action and recurrence—not merely a low incident count. The strongest counter-case is therefore not that oversight criteria are unnecessary. It is that badly designed criteria could make the information worse.

DOE's documented 2025 and 2026 milestones matter too. Waste was shipped, contaminated equipment stabilised and water protection work continued. These outputs should be measured separately from whether the internal system is sufficiently reliable. A responsible editorial judgment can credit those outcomes while asking why headquarters could not show a functioning risk-based review schedule.

Finally, the August 2026 audit itself is bounded: three non-representative contracts, historical evaluations and management interviews. It does not establish the safety status of every DOE site, the amount of any avoidable cleanup cost, the legal liability of any contractor, or the current completion status of DOE's four corrective commitments.

What would change the finding

The decisive records are not another speech about efficiency. They are a signed, dated definition of CAS effectiveness; field-office assessments using it; standardised metrics with documented exceptions; implemented headquarters oversight schedules; and contract clauses or performance plans showing what happens when a contractor misclassifies significant issues or fails to correct them. Fee-determination letters and actual payment data could establish whether financial incentives changed. Incident investigations could establish whether particular weaknesses caused harm. Congressional hearings and appropriations could show whether oversight staffing has been restored sufficiently to validate contractor reports.

If those records demonstrate robust independent checks, prompt correction and meaningful contractual consequences, the central criticism should be revised. If they remain absent while contractors continue to be labelled effective, the gap between a public assurance and the evidence behind it will persist.

Conclusion: a system cannot certify itself by adjective

The August audit did not uncover a missing $840 billion, a proven criminal conspiracy or a universally unsafe cleanup industry. It uncovered a more precise and actionable accountability problem: federal managers could not show what made a contractor's assurance system effective, headquarters lacked a consistent way to use site information, and the reviewed contracts did not establish clear consequences for that system's poor performance. The earlier workforce audit raises the stakes without proving causation. Chris Wright and Timothy Walsh did not originate every defect, but they now possess authority to make the federal test explicit, auditable and enforceable. The public-interest measure is not whether the department says it is committed to safety. It is whether the next independent reviewer can reproduce the rating and trace it to a real decision.

Factrail context and separate accountability mechanisms

The Department of Energy works under the Donald Trump administration (opens in a new tab), but the GAO case histories arose over several years and cannot be assigned solely to current leaders. Effective cleanup and worker safety are distinct from the availability problem tracked under energy security. Weak contractor-assurance rules are relevant to institutional-corruption safeguards (opens in a new tab) as a governance category, not proof of bribery or unearned fees.

Source register and limitations

The documentary backbone is GAO-26-107850, DOE Contracting (26 August 2026) (opens in a new tab), cross-checked against GAO-26-108674, Nuclear Waste Cleanup Workforce (19 May 2026) (opens in a new tab) and DOE Office of Enterprise Assessments, Contractors' Management of Safety Issues (24 April 2024) (opens in a new tab). Contract identities and ceilings are cross-checked with DOE's Hanford award notice (opens in a new tab), Jacobs' Idaho award statement (opens in a new tab), and Idaho Environmental Coalition's current description (opens in a new tab). The counter-record includes DOE's Los Alamos performance report (opens in a new tab) and October 2026 Hanford progress report (opens in a new tab). AP's confirmation reporting (opens in a new tab) independently documents Wright's appointment context. This article did not obtain unredacted fee decisions, original contractor assurance logs, fresh comment from the named contractors, or proof of implemented corrective actions after the audit. No right of reply is represented as having been requested. Factrail context links and Commons attribution were added in the 1 November 2026 editorial review; the direct image response remains to be checked.

Sources

1. GAO — DOE Contracting: Risk-Informed Oversight and Clearer Expectations for Assurance Systems (26 August 2026) (opens in a new tab)
2. GAO — Nuclear Waste Cleanup Workforce (19 May 2026) (opens in a new tab)
3. Department of Energy — Independent Assessment of Contractors' Management of Safety Issues (24 April 2024) (opens in a new tab)
4. Department of Energy — Hanford Central Plateau Cleanup Contract Award (opens in a new tab)
5. Jacobs — Idaho Cleanup Contract Award (opens in a new tab)
6. Idaho Environmental Coalition — Contractor and Project Information (opens in a new tab)
7. Department of Energy — Los Alamos Legacy Cleanup Safety Progress (2025) (opens in a new tab)
8. Department of Energy — Hanford Glovebox Stabilization Report (October 2026) (opens in a new tab)
9. Associated Press — Chris Wright Appointment Context (2025) (opens in a new tab)