By Lucy Wang
In July 1945, the world entered the nuclear age. The Trinity test marked the first successful detonation of a nuclear weapon. In the decades that followed, nuclear technology powered submarines, generated electricity for millions of households, and ushered in the commercial nuclear power industry. Nuclear engineering became one of the defining scientific disciplines of the twentieth century. Yet over the following decades, nuclear development and investment slowed due to high construction costs, lengthy regulatory processes, and public backlash following nuclear accidents like Three Mile Island, Chornobyl, and Fukushima.
By the early twenty-first century, many see the nuclear industry as mature, with limited room for future growth. But the development of artificial intelligence is changing that perception. AI is transforming nuclear engineering practices, while nuclear energy is emerging as a solution to one of AI’s greatest challenges: its enormous demand for electricity.
According to Igor Jovanovic, a professor of nuclear engineering at UC Berkeley and a faculty affiliate of Lawrence Berkeley National Laboratory, “AI is accelerating the discovery of advanced radiation-resistant materials, optimizing fission reactor designs and predictive maintenance, and enabling real-time plasma control in fusion energy systems.” AI has the potential to make nuclear technologies more efficient and reliable, and at the same time, nuclear energy stands to be one of the most promising energy sources to power the growth of AI. Future breakthroughs in AI depend on a few important resources: efficient algorithms, advanced semiconductor chips, data centers that provide computing capacity, and electricity to power the physical infrastructure. The US continues to lead in advanced chips, while China is rapidly narrowing the gap in developing the most advanced AI models.
AI is a digital technology, but its greatest constraints are physical. AI models are trained in data centers, which are large facilities filled with servers and networking equipment that are the foundation of cloud computing, e-commerce, streaming services, and AI. The US has a staggering 4,423 data centers, which is more than the next ten countries combined. But building the computing hardware is only half of the challenge; powering it is much more difficult. Every graphics processing unit (GPU) inside a data center requires electricity. Over 4 percent of all the energy in the US now goes toward data centers; that number is expected to rise to 12 percent by 2028. Between now and 2030, “data centers account for nearly half of electricity demand growth” in the US. China currently has the greatest capacity in the world to produce electricity by burning coal and expanding renewable energy sources like solar and wind. To meet this unprecedented energy demand, the US is pursuing a range of energy solutions including solar and wind while also reinvesting in nuclear power.
“We are at one of the most exciting inflection points in the history of nuclear engineering,” says Jovanovic. “There has rarely been a time with greater opportunity to make an impact on global nuclear energy technology and policy.” This renewed demand for reliable electricity for AI and rapid private investment in nuclear technologies are creating opportunities across industry, academia, national laboratories, startups, and government. For students considering the field, there has rarely been a better time to enter the profession. The International Atomic Energy Agency (IAEA) argues that nuclear power is uniquely positioned to support AI because it provides stable, low-carbon baseload electricity independent of weather conditions, which is well suited to energy-intensive computing infrastructure.
The US Department of Energy (DOE) has called for a nuclear renaissance. Through its “Manhattan Project 2.0” initiative, the DOE proposes repurposing former Cold War nuclear sites into centers for advanced reactors, domestic nuclear fuel production, and other nuclear technologies needed to power the next generation of AI. The renewed interest in nuclear energy is driven by tech companies leading the AI race that are investing in nuclear power. Microsoft signed a 20-year agreement to purchase the entire output of the restarted Three Mile Island reactor; Google contracted electricity from seven planned Small Modular Reactors (SMRs); and Amazon invested $500 million in X-energy reactors while securing power from existing nuclear plants.
This transformation creates enormous opportunities for the next generation of nuclear engineers. Nuclear engineering offers a diverse range of career paths including “advanced fission reactor development, fusion energy startups, space power and propulsion, nuclear medicine, national security, radiation instrumentation, and policy,” says Jovanovic. However, success in the field will require more than technical expertise. According to Jovanovic, nuclear energy projects increasingly operate at the intersection of technology, finance, regulatory policy, public perception, and supply chain logistics. The next generation of nuclear engineers will have to engage in cross-disciplinary collaboration. For many considering the field of nuclear engineering, Jovanovic encourages students to gain practical research and industry experience while learning how new technologies move “from proof of concept through manufacturing, licensing, economic viability, and public acceptance.” The work of nuclear engineers will be central to one of the world’s fastest-moving technological revolutions. Eighty years after the Trinity test introduced the world to the atom, artificial intelligence has given nuclear energy a renewed purpose.
But the future of AI will depend on more than engineers alone. Building the infrastructure and energy system that power AI requires enormous investment in talent and technology, but all of it relies on public trust. Data centers, transmission lines, and nuclear reactors cannot be built without the support of local communities, where projects face intense opposition over concerns about land use, zoning, and environmental impacts. Maintaining US leadership in AI has become a national security imperative, and national security depends not only on military or technological strength, but also on an informed public that understands and believes in what it is defending. As the United States advances in the AI race, bringing the American people along may prove just as important as maintaining its technological edge.