What NNSA’s FY 2025 Performance Evaluation Report Tells Us About Lawrence Livermore Lab
Every year, the National Nuclear Security Administration (NNSA) quietly publishes a report card for Lawrence Livermore National Laboratory (LLNL), managed by Lawrence Livermore National Security, LLC (LLNS), and if they get good grades, they get tens of millions in tax-payer funded contractor bonuses. The FY 2025 Performance Evaluation Report (PER), covering October 2024 through September 2025, tells us a great deal about what’s happening inside the Lab — and what risks the surrounding community may be bearing as a result.
This year, LLNS earned an overall “Excellent” rating and 91 percent of its available award fee — roughly $46.6 million out of a possible $51.1 million. However, behind those numbers lie troubling details that demand public scrutiny.
Plutonium Pit Work: Livermore’s Expanding Role
The most consequential activity highlighted in this year’s PER — and the one that most directly implicates community health and safety — is Livermore Lab’s deepening involvement in the nation’s plutonium pit production program.
Plutonium pits are the fissile cores of nuclear warheads. The NNSA is planning on producing new plutonium pits, which it describes as “central to the United States’ nuclear weapons modernization drive” on new production lines at both Los Alamos National Laboratory (LANL) in NM, and Savannah River Site (SRS) in SC. This plan is the subject of Tri-Valley CAREs successful litigation (along with our colleague groups Nuclear Watch New Mexico and Savannah River Site Watch) and the court ordered Programmatic Environmental Impact Statement that is set for release in May 2026.
There is no demonstrated need for new plutonium pits to ensure the safety or reliability of the existing US nuclear weapons stockpile. In fact, these new pits will be for entirely new nuclear weapons, specifically the W87-1 warhead for the new Sentinel Intercontinental Ballistic Missile (ICBM) for which LLNL is the lead design lab. This warhead will take up at least the first decade of pit production at LANL.
It should also be noted that this increased plutonium pit work is also the driver for LLNL’s proposed “Enhanced Plutonium Utilization,” which is steamrolling ahead without public involvement in the decision-making process.
The PER specifically documents a significant and wide-ranging set of pit-related activities at LLNL in FY2025. It completed nine “Qualification and Engineering Releases” in support of pit production, engaged in over 120 “Product Realization Teams,” and successfully led the “Pit Product Realization Team (PRT)”. The Lab also benchmarked operations and infrastructure with LANL’s plutonium facility — where actual pit manufacturing is already ramping up — exploring topics including safety basis activities for a new LANL plutonium facility, Material Control and Accountability (MC&A), and Radiation Protection. The Lab also performed “testing on the safe storage of pits at Pantex” Plant in Texas.
Perhaps most striking, the report notes an 18-month pilot project between the Livermore Field Office and LLNS to demonstrate “the practicality and feasibility of managing risks cross-functionally in multiple projects to include the Superblock Facility complex.” The Superblock is Livermore’s plutonium and highly enriched uranium facility — the most sensitive and dangerous building on the site. This pilot is intended to enable “faster decision making and execution” and “streamline operations.” Traditionally, faster decision making and operations in a facility that handles weapons-grade nuclear material is not a safety feature, but rather a safety risk. The community deserves to know exactly what changes to operations and decision making are being made in order to speed up these dangerous processes.
The picture that emerges is of a Lab whose plutonium-related footprint is speeding up, expanding, evolving, diversifying, and deepening its ties to the broader nuclear weapons production complex.
Testing: Hydrotests, NIF, and Weapons-Grade Plutonium in the Laser Bay
While the United States has maintained a moratorium on nuclear explosive testing since 1992, Livermore conducts a range of non-nuclear and subcritical experiments to gather data on how nuclear weapons perform. The FY 2025 report documents significant activity on multiple fronts.
The report confirms that LLNS completed “pressing and machining of test articles for upcoming hydrodynamic tests” — high-explosive experiments that compress nuclear materials to near-detonation conditions — planned for FY 2026. Details about these experiments are heavily classified, and the scope, frequency, and risk profile of tests conducted remain largely invisible to the public. Tri-Valley CAREs aims to seek further information with a FOIA request.
The National Ignition Facility is frequently promoted for its fusion energy potential, but the PER is a useful reminder that NIF’s primary mission is weapons science. This year, the facility conducted nine successful fusion ignition experiments, with the eighth setting a new yield record of 8.6 megajoules. These experiments generate data directly relevant to nuclear weapons physics. Most strikingly, the report confirms that LLNS executed a NIF shot using a weapons-grade plutonium sample — bringing actual weapons material into an ignition facility and raising questions about nuclear material handling and safety that the public has a right to ask.
The Lab also conducted the “Nob Hill” subcritical experiment at the Nevada National Security Site, and references an unnamed classified experiment. Taken together, Livermore is aggressively using every available tool — high explosives, subcritical experiments, and laser-driven ignition — to advance nuclear weapons science in the absence of underground testing.
Artificial Intelligence: From Infrastructure Decisions to Shaping Federal Policy
Artificial Intelligence (AI) is woven throughout the FY 2025 PER in ways that deserve careful scrutiny. The Lab is deploying AI not just for basic research, but directly in support of nuclear weapons design, infrastructure management, and national security policy — and doing so rapidly, with limited public oversight.
The report documents the deployment of a dedicated AI system on the Lab’s restricted zone network, with plans to eventually transition it to the classified network — meaning AI infrastructure operating in proximity to classified nuclear weapons data, with no public accountability framework described. The report also introduces “Bernie AI,” a new machine learning tool built to support infrastructure decision-making using NNSA data. While infrastructure planning may sound mundane, NNSA’s infrastructure encompasses nuclear weapons production facilities and contaminated buildings handling weapons-grade materials. Automating decisions about that infrastructure raises legitimate questions about whether human judgment and accountability are being adequately preserved.
More broadly, the PER states that LLNS subject matter experts “helped shape a methodology that has transformed NNSA’s understanding of AI-related risks and in identifying pathways for the broader US Government to adopt and integrate AI/Large Language Model technologies.” Livermore scientists are not just using AI — they are influencing how the entire federal government thinks about and deploys it in national security contexts. That is an extraordinary level of institutional influence over a technology developing faster than any governance framework can currently track.
The Lab is also applying AI directly to weapons science — using machine learning in material response and plasma models that simulate nuclear weapon behavior, and leveraging it to screen energetic binder candidates for plastic-bonded explosives. Underlying all of this is El Capitan, the world’s most powerful supercomputer, which transitioned to classified network production status at Livermore in September 2025. The PER is explicit that El Capitan exists to “ensure the safety, security and reliability of the nation’s nuclear deterrent without nuclear testing.” As AI accelerates what is possible on systems like El Capitan, the line between maintaining the existing stockpile and developing new nuclear capabilities becomes harder to draw — and the public deserves to be part of that conversation.
Safeguards and Security: Persistent Failures the Lab Can’t Fix
While LLNS touts its weapons accomplishments, the PER provides a surface-level glimpse into some ongoing security failures — particularly around plutonium and nuclear material accountability. Under Goal 4 (Mission Enablement), the Lab received only a “Very Good” rating, and the security section is worth a close review.
The report documents that the Safeguards and Security program suffered from “inconsistent analysis of recurring and systemic issues, prolonged corrective actions, and slow escalation of Incident of Security Concern (IOSC)” — problems that have “weakened” the Lab’s “ability to detect and mitigate emerging risks.” These failures were compounded by data integrity problems, including incomplete records and misalignments in the Issues Tracking System.
On Material Control and Accountability — the program responsible for tracking nuclear materials like plutonium — the report acknowledges that “issues continue to persist” even after the Lab completed a recent inventory and made management improvements. The Lab still needs to achieve full compliance with its MC&A Plan and DOE Order 474.2A. In plain language: the Lab is not fully meeting federal standards for tracking its weapons-grade nuclear material, while simultaneously proposing that it be allowed to increase its plutonium throughput without any public scrutiny or input.
There were also unresolved Personnel Security Deficiencies carried over from FY 2023, hiring restrictions in key security roles, and unresolved findings from a FY 2024 Program Management survey. Hiring restrictions hindering security improvements at a nuclear weapons facility — at the same time that its plutonium workload is expanding — should be alarming. These are precisely the kinds of compounding failures that Tri-Valley CAREs has long warned can precede more serious incidents.
Cost Accounting Problems with Real Consequences
The PER also reveals financial management failures with direct programmatic consequences. The report describes how LLNS submitted an inaccurate cost estimate for equipment removal from Building 435, forcing NNSA to reduce available FY 2026 funding by $2.9 million. Additionally, the Lab submitted inadequate documents for the decommissioning of Building 212 — a process-contaminated facility. Properly deactivating contaminated buildings is critical to protecting workers, the public, and the environment. When planning for that work is based on flawed estimates and incomplete submissions, the consequences extend well beyond budget lines.
The Current Administration’s Heavy Hand
The Trump Administration continues to bear down its heavy hand over the facility as the PER stated that “LLNS’s senior leadership and Human Resource (HR) staff took prompt actions to comply with new Executive Orders by addressing DEI program staffing, policies, and programs; and gender ideology policies and practices as required. Senior leadership also focused on removing external references to climate change and related topics.” Tri-Valley CAREs finds this objective in the PER to be outrageous, signaling the administration is inserting its own personal beliefs and motives into operations. In regards to removing references to climate change, this is incredibly worrisome as the administration has been on a crusade to slash funding for renewable energy for the facility. This makes it clear that the primary purpose of the lab is for nuclear weapons operations.
The Bottom Line
The FY 2025 PER presents a Lab that is expanding its nuclear weapons work, deepening its involvement in plutonium operations, integrating AI into weapons-sensitive systems with minimal transparency, and earning high marks from its federal overseers — while simultaneously failing to meet basic requirements for nuclear material tracking and security assurance. The report’s overall “Excellent” rating should not obscure the seriousness of the problems documented within it.
Livermore Lab sits in the middle of a densely populated region in the San Francisco Bay Area. The decisions made inside the Superblock — about plutonium, about security, about how quickly to streamline operations — have consequences for the hundreds of thousands of people who live and work nearby. They deserve a seat at the table, not just a press release.
Tri-Valley CAREs will continue to monitor the Lab’s activities, pursue public records, and report back to our community.