Ampera, a US nuclear startup, has unveiled what it claims is the first 3D-printed nuclear reactor module, targeting AI data centers and defense applications with a thorium-based design that could deliver power for three decades without refueling. The prototype features a fully 3D-printed silicon carbide reactor core and pressure vessel, marking a significant shift toward factory-built, mass-produced nuclear energy using advanced manufacturing techniques.

Why Is 3D Printing Nuclear Reactors Such a Big Deal?

Traditional nuclear reactor construction relies on conventional manufacturing methods that are slow, expensive, and difficult to scale. Ampera's approach uses additive manufacturing to create a spherical monolithic gyroid core, a complex geometric shape that provides massive surface area relative to its volume, making it ideal for heat transfer. This shape would be nearly impossible to produce using conventional methods, but 3D printing makes it feasible.

The innovation addresses a critical bottleneck in the AI industry: power. Data centers training large language models consume enormous amounts of electricity, and grid infrastructure often cannot keep pace with demand. By manufacturing reactors in factories rather than building them on-site, companies like Ampera aim to accelerate deployment timelines significantly.

"This next-generation nuclear core and pressure vessel sets the foundation for factory-built, mass-produced nuclear energy," said Brian Matthews, founder and CEO of Ampera. "The advanced technology and additive manufacturing used demonstrate a clear commercial path for new nuclear technology coming to market in an accelerated manner."

Brian Matthews, Founder and CEO at Ampera

How Does Ampera's Thorium Reactor Design Work?

Ampera is developing a subcritical, solid-state, factory-built thorium-based nuclear reactor. Subcritical means the fuel cannot sustain a nuclear chain reaction on its own, which prevents dangerous runaway power excursions. The reactor uses TRISO (tristructural isotropic) particles, consisting of a fuel kernel containing thorium surrounded by multiple ceramic and carbon layers.

Thorium-232 is not fissile on its own. After absorbing a neutron, it decays through thorium-233 and protactinium-233 into fissile uranium-233. This process requires a separate source of neutrons, and Ampera says its design features a proprietary neutron driver to provide a stable external neutron source to start and sustain operation. The company is keeping the specifics of this neutron driver confidential for now.

Steps to Understanding Ampera's Path to Market

• Fuel Supply Strategy: In June, Ampera announced it had established an Australian subsidiary to secure thorium supplies and plans to produce TRISO thorium fuel kernels itself in the United States, ensuring ample access and minimizing price volatility risk.
• Power Output Specifications: The planned systems will provide either 15 or 30 megawatts of electricity depending on configuration, enough to supply a typical data center, with larger configurations planned for future deployment.
• Operational Lifespan: The 3D-printed silicon carbide core is designed to operate for up to 30 years without refueling, significantly reducing maintenance costs and operational complexity compared to conventional reactors.
• Deployment Timeline: Ampera expects the power generation portion of the system to be available as early as 2027, with the nuclear module becoming available to customers around 2030 based on regulatory approval.

How Does This Fit Into the Broader Nuclear-AI Boom?

Ampera's announcement comes as the nuclear industry experiences unprecedented momentum. In early July 2026, multiple nuclear startups achieved criticality, the state in which a reactor sustains a stable chain reaction, meeting a deadline set by President Donald Trump in May 2025. This milestone demonstrates that the nuclear sector is moving beyond theoretical concepts toward practical, deployable systems.

Several startups have already demonstrated significant progress. Valar Atomics achieved criticality twice, including a June test where it ramped its Ward250 reactor to full power and generated 100 kilowatts of electricity, enough to power an Nvidia Blackwell processor. The company subsequently announced a deal with Nvidia to build a 30-megawatt nuclear data center in Utah. Antares Energy's Mark-0 reactor became the first privately funded non-light-water reactor to achieve criticality at Idaho National Laboratory in four decades. Deployable Energy, founded just a year ago, created a 1-megawatt shipping-container reactor called the Unity Nuclear Battery and demonstrated its portability by driving the reactor core to Idaho Falls in a Ford F-150.

Beyond these recent achievements, other companies are pursuing ambitious plans. Kairos Power is simultaneously erecting two pilot plants near Oak Ridge National Laboratory in Tennessee, with its larger Hermes 2 reactor rated at 50 megawatts and set to sell electricity to the Tennessee Valley Authority to power nearby Google data centers, potentially as early as 2030. Amazon has committed to buying 5 gigawatts of reactors from X-Energy, which raised $1 billion in an April IPO.

The convergence of advanced manufacturing, regulatory support, and massive corporate demand for clean power is reshaping the energy landscape. Ampera's 3D-printed approach represents one strategy among many to accelerate nuclear deployment, but the underlying trend is clear: the AI industry's insatiable appetite for electricity is reviving nuclear power as a cornerstone of future infrastructure.

However, significant challenges remain. Most of these new reactor designs require TRISO fuel pellets that are not yet widely available, and none of the new reactors have yet proven they can operate at full power for extended periods or generate electricity in sufficient quantities to reliably power data centers at scale. The next few years will determine whether these startups can translate their engineering achievements into commercially viable, revenue-generating systems.