As the global nuclear resurgence increases pressure to deliver projects on schedule and within budget, AI tools are gaining traction as ways to manage the extreme complexity of both new construction and decommissioning, writes ALICE Technologies Director of Americas & APAC Kevin Fuller. These platforms help teams simulate build strategies, revise schedules in real time and model supply chain risks before they cascade into delays. Although AI cannot eliminate structural challenges, it gives project developers greater visibility into risks and trade-offs, supporting more predictable project delivery, Fuller writes.

A team of researchers from the University of California, Los Angeles has developed an electroplating technique that replaces complex crystal-growth methods for thorium-based nuclear clocks and uses only trace amounts of thorium-229, one of Earth's scarcest isotopes. Researchers found that only surface nuclei need to be excited, meaning the host material no longer has to be transparent. The breakthrough could make thorium-based nuclear clocks cheaper and easier to build, opening the door to GPS-independent navigation and other high-precision timing applications.

President Donald Trump's vision for a "nuclear energy renaissance" aims to quadruple US nuclear power by 2050, but big challenges remain. Only two new reactors have come online in the past decade, both at the Alvin W. Vogtle Electric Generating Plant, which was $17 billion over budget and seven years late. Trump has promised $80 billion to build reactors, with early projects receiving subsidies, but construction often faces delays and cost overruns because each plant is treated as a first-of-a-kind project. Small modular reactors are seen as a more flexible option, though they may be more expensive per megawatt-hour.

The Electric Power Research Institute tested a variety of large language models, including GPT-5, Gemini 2.5 Pro, and Claude Sonnet, on issues facing US power systems and found that they performed poorly on open-ended engineering questions. GPT-5 managed a 63% success rate on the most challenging questions, while other models scored between 46% and 59%, highlighting reliability concerns.

The Wisconsin Legislature is considering a bill that would provide state tax credits to encourage the development of new nuclear power plants. The credits would be available for plants that begin operating in 2030 or later, as well as for existing plants that are relicensed by federal regulators. By helping offset the significant startup costs of nuclear projects, lawmakers hope to make nuclear energy more economically viable in the state.

The Environmental Protection Agency has revised its climate change webpage to remove references to coal, oil and natural gas as drivers of global warming, instead emphasizing only natural phenomena such as solar activity and volcanic eruptions. Scientists and former agency officials say the change misrepresents consensus science and misleads the public, but an agency spokesperson defended the changes as consistent with the current administration's priorities and rejected accusations of misinformation.