Venture Bytes #120: Nuclear Power Key to Industrial Automation

Nuclear Power Key to Industrial Automation

Humanoid robots are marching into factories faster than the grid can keep up. Andwhile automation is scaling like software, power isn’t. This asymmetry is setting up amajor supply-demand mismatch, creating a breakout opportunity for nuclear startups,particularly small modular reactor (SMR) startups.President Donald Trump’s reshoring agenda has created the perfect storm for investmentsin next-generation energy and robotics. But tariffs alone won’t spark a manufacturingrenaissance. Labor costs in the US remain high, and skilled workers are scarce. As of2023, only 9.7% of US private sector workers are in manufacturing, down from 31% in1970, according to the Economic Innovation Group.Therefore, the real catalyst for reshoring will be the rise of humanoid robotics, whichis poised to redefine how factories operate. A CNBC survey found that 81% of thecompanies’ reshoring production plans will rely on automation over human labor.But this automation wave comes with a catch: the need for abundant energy. Accordingto the Association for Advanced Automation, a single 210 kg payload robot can consumean estimated 60,000 kWh over its lifecycle, depending on motion patterns, idle time, andstoppages. This added load compounds the strain on power grids already under pressurefrom energy-hungry AI data centers, creating a strategic opportunity for energy-focusedstartups.

According to Manhattan Venture Research estimates, the U.S. could have over 1.2 million robots operating in factories by 2027, if it aims to catch up with global leaders in robot density like Korea, Japan, and Germany. If each robot works 16-hour days across 260 business days, the annual additional power burden could hit 1,290 GWh—roughly equivalent to the annual power needs of Austin, Texas.

This new robotic industrial layer will run in parallel to another energy-intensive layer: AI infrastructure. Hyperscale data centers are already reshaping regional power markets. Now, factories are next. The grid wasn’t built for this. Legacy fossil plants can't scale fast enough. Solar and wind are intermittent. Batteries are expensive. What’s needed is clean, 24/7, modular power that scales with demand. That’s the precise niche SMRs are built to fill.

This creates a massive opportunity for clean energy startups, particularly in nuclear fusion, which promises clean, reliable power to fuel the robotics boom. SMRs fit the bill perfectly. They are compact nuclear power plants designed to produce electricity more safely, flexibly, and affordably than traditional large reactors. SMRs can generate carbon-free electricity, reducing the environmental footprint of the robotics revolution. Their ability to provide a reliable, 24/7 power supply is crucial, unlike intermittent renewable energy sources such as wind or solar. Furthermore, SMRs’ modular nature allows for phased deployment, scaling in tandem with the evolving energy requirements of expanding robotics adoption.

Favorable policy momentum is accelerating SMR adoption. On May 23, 2025, President Trump issued executive orders to fast-track the deployment of advanced nuclear reactors and expand U.S. nuclear capacity from approximately 100 GW to 400 GW by 2050. This marks one of the most aggressive nuclear infrastructure pushes in decades, aimed at securing long-term energy independence and supporting emerging industrial loads like robotics and AI. The global market for SMRs is expected to grow from $5.8 billion in 2022 to $13.4 billion by 2032, driven by rising demand for clean, resilient baseload power that can scale with 21st-century manufacturing systems.

Startups such as X-Energy, Last Energy, and Terrapower are well-positioned to leverage this burgeoning market. In October 2024, X-Energy partnered with Amazon to deploy over 5GW of new power projects across the US by 2039, marking the largest commercial deployment of SMRs to date. 

Last Energy, reflecting investor confidence, experienced a 382% valuation increase in its most recent funding round and had secured commercial agreements for 80 SMRs as of August 2024, up from 34 in 2023. Additionally, in January 2025, Terrapower signed an MoU with Sabey Data Centers (SDC), one of the largest private data center providers globally, to integrate Terrapower's SMRs into SDC's existing and future data center operations.

The real winners of the next industrial wave will be humanoid robots—and the SMRs powering them. Start-ups like Figure AI and Agility Robotics are leading the automation layer, while X-Energy, Last Energy, and Terrapower are building the power stack beneath it.

Rare Earths Aren’t Rare, Refining is the Catch

Rare earth elements have recently emerged as critical levers of geopolitical influence and technological dominance. From electric vehicles (EVs) and wind turbines to precision-guided missiles and advanced computing, these materials are embedded in nearly every industrial system that defines modern power, both economic and military. But just how rare are they? And more importantly, what are the viable alternatives for the US as it seeks to reduce strategic dependence on China?

The recent trade tensions with China have offered a sobering wake-up call that the US cannot afford to rely on a single geopolitical rival for access to inputs so foundational to its industrial and defense base. The urgency becomes even clearer when considering the scale of future demand. Industry analyst Jack Lifton estimates that if the US automotive sector shifts entirely to EVs, producing 17 million cars annually (which is the average), it would require approximately 40,000 tons of rare earth permanent magnets per year. And EVs are only one piece of the equation. Rare earths are also essential to critical technologies such as high-efficiency motors, missile guidance systems, wind energy, and even everyday items like smartphones and hard drives.

If demand is a certainty, then availability deserves a closer look. The label “rare earth elements” is, in fact, a misnomer. These elements — comprising the 15 lanthanides, along with scandium and yttrium — are relatively abundant in the Earth’s crust. Elon Musk recently pointed this out on X, noting that the challenge isn’t scarcity, but access and processing. Cerium, for example, ranks as the 25th most abundant element, with a crustal concentration of 68 parts per million — higher than that of copper. Even the least abundant REEs, such as thulium, are present at about 0.5 parts per million.

But what Musk wrote was an oversimplification of the challenge. The "rarity" of these elements stems from their dispersed nature, rarely forming concentrated deposits that are economically viable to mine. Extracting and refining rare earth elements requires processing vast amounts of ore, a costly and complex endeavor. This is where China has executed a long-game strategy. Since Deng Xiaoping’s 1992 declaration — “The Middle East has oil, China has rare earths” — the country has methodically built dominance across the entire value chain.  

If China’s rare earth dominance is the product of long-term industrial planning, the US response will likely be driven by technological ingenuity and market-led innovation. That’s where a new class of US-based startups is beginning to play a crucial role, presenting compelling investment opportunities across the value chain. Instead of relying solely on new mining projects that take 10–15 years to reach production, these companies are creating vertically integrated, closed-loop systems that can be scaled far more quickly.

Massachusetts-based Phoenix Tailings exemplifies this shift. The company raised $76.8 million in Series B funding recently and is rapidly scaling toward 500 tons of annual rare earth metal production capacity. What sets Phoenix apart is its zero-waste, environmentally safe refining process, which eliminates the need for harmful solvents and avoids radioactive byproducts, two major barriers to domestic processing. Its current focus is on high-purity outputs for defense, medical device, and automotive applications, offering a secure, U.S.-based source for sectors where supply disruptions are not just costly but strategically dangerous.

Texas-based Noveon Magnetics is filling a different but equally urgent gap with permanent magnet manufacturing. As the only US producer of sintered neodymium-iron-boron magnets, Noveon is now scaling up to meet surging Western demand. Its production capacity has doubled to 2,000 tons annually by 2024, and it uses a patented “magnet-to-magnet” recycling process that is over 90% more energy-efficient than traditional oxide-based production. This closed-loop system gives Noveon a cost and time advantage while reducing exposure to virgin raw material imports.

Another key start-up is Niron Magnetics, which is reimagining the fundamentals of magnet chemistry. Niron develops rare-earth-free permanent magnets using iron and nitrogen, a breakthrough that could dramatically reduce US dependence on rare earths altogether. Niron is backed by major industry players including GM, Stellantis, and Samsung, underscoring broad confidence in its commercial viability. The company is planning a 150,000-square-foot manufacturing facility in Sartell, Minnesota, targeting 1,500 tons of annual production by 2026–2027.

Another strong player is Kobold Metals, which is transforming the frontier of discovery. While most attention focuses on processing and production, exploration remains the most under-optimized segment of the critical mineral value chain. Kobold uses AI and machine learning to accelerate mineral discovery, combining geoscience, satellite data, and proprietary algorithms to model subterranean geology. Kobold currently has over 60 exploration projects across three continents, targeting not just rare earths but a broader suite of critical minerals essential to clean energy, semiconductors, and defense technologies.

More broadly, these startups represent a growing ecosystem fueled by rising VC investment, an expanding pipeline of innovation, and growing federal alignment. As the global rare earth metals market is projected to hit $16.1 billion by 2034 from $6.2 billion in 2024, driven by electrification and national security demand, these start-ups sit at the convergence of industrial necessity and policy urgency.