DARPA’s Drone Power Gamble Stuns Scientists

The most consequential story in DARPA’s “Rads to Watts” effort is not that nuclear waste might one day power next‑generation drones, but that the Pentagon is deliberately trying to leap from milliwatt‑class nuclear batteries to kilowatt‑scale solid‑state power in one stroke—against a backdrop of decades of stalled promises and hard physical limits.

Key Points

DARPA’s Rads to Watts program is a formal, multi‑year push to develop high‑power radiovoltaic devices that turn nuclear radiation directly into electricity, with an explicit kilowatt‑scale vision.
The program is structured around durable “unit cells” and competitive endurance testing, and includes contracts like Avalanche Energy’s >10 W/kg alpha‑voltaic battery for laptop‑class loads.
Public reporting links Rads to Watts to nuclear‑waste‑derived cells for long‑endurance drones, but specific claims about 30‑year lifetimes, particular isotopes, and named projects are not corroborated in primary DARPA documents.
Mainstream scientific consensus and past experience with nuclear batteries point to severe power‑density and regulatory constraints, suggesting this is an ambitious research program rather than a near‑term operational drone power source.

From Milliwatts to Kilowatts: What Rads to Watts Is Trying to Do

DARPA’s Rads to Watts program is not a speculative rumor; it is a formally announced Defense Sciences Office effort with a special notice and program page that lay out a demanding technical agenda. At its core, Rads to Watts aims to develop radiovoltaics—solid‑state devices that convert nuclear radiation directly into electrical power, bypassing conventional heat engines and turbines used in radioisotope generators. The official language is strikingly bold: the “ultimate vision” is radiovoltaic systems that take “high‑power nuclear radiation” and deliver electricity at the kilowatt level, operating as long‑lived, unattended power sources in domains where resupply is impossible. Those domains range from deep space and lunar nights to “dark, remote places on Earth,” which naturally includes autonomous systems such as drones.

The challenge is that current radiovoltaic technologies have effectively plateaued at milliwatt outputs per unit cell. Traditional betavoltaic or alphavoltaic devices can run ultra‑low‑power sensors for decades, but scaling those designs to kilowatt levels would require impractically large quantities of radioisotope material. Rads to Watts is explicitly framed as an attempt to break through that plateau by forcing performers to abandon “typical low‑power radiovoltaic architectures” and invent new materials and device structures that can survive high radiation fluence—particle flux over area—without rapid degradation.

How the Program Is Structured: Unit Cells, Time Capsules, and Power Contests

DARPA has gone to unusual lengths to nail down the technical vocabulary and competitive structure of Rads to Watts. The program solicitation defines a radiovoltaic “unit cell” as the smallest viable building block, scalable for a given power density and comprising three fundamental regions: a charge generation region (CGR), where incident radiation creates electron–hole pairs or charges; a charge collection region (CCR), which gathers and routes that charge; and a source region, where the radiation is emitted. This tri‑part architecture is not merely definitional—it is the kernel around which device designs must be optimized for both output and survivability.

To stress these unit cells well beyond the benign environments typical of low‑power nuclear batteries, DARPA has designed the program as a “power contest.” Teams compete to maximize power density and durability under punishing radiation from a U.S. Army Research Laboratory linear accelerator, then face a second down‑select based on samples sequestered in “time capsules” and exposed to even higher flux. Concepts that survive both stages may enter a bonus period where they are integrated into radioisotope‑fueled systems tailored to high‑radiation regions of space or remote terrestrial sites. This contest structure reflects the program’s central technical bet: that the right combination of materials, geometry, and source choice can maintain acceptable performance under conditions that currently destroy semiconductor devices.

Contracts and Prototypes: Avalanche Energy and the >10 W/kg Target

Beyond general notices, Rads to Watts already has named performers and specific performance targets. Avalanche Energy, a Seattle‑based startup, has secured a multi‑million‑dollar DARPA contract under Rads to Watts to develop micro‑fabricated alphavoltaic cells that convert alpha particles from radioisotopes directly into electricity. The contract’s goal is ambitious but quantitatively modest compared to the kilowatt vision: a device achieving more than 10 watts per kilogram, roughly a 10‑pound unit capable of powering a laptop‑class system continuously for several months. For context, radioisotope thermoelectric generators on Mars rovers deliver only a few watts per kilogram.

This >10 W/kg target matters for two reasons. First, it is far above the microwatt‑to‑milliwatt regime of most nuclear batteries discussed in popular videos and academic work, which typically aim at tiny sensors or medical implants rather than sustained computing loads. Second, it is still orders of magnitude below what a full‑scale drone propulsion system requires; the immediate intent is to power onboard electronics and observation payloads, not replace all propulsion power. Avalanche’s design will be validated using both particle accelerators and active radioisotopes, and its eventual applications are explicitly framed around autonomous military systems, remote bases, subsea vehicles, and space platforms, rather than consumer gadgets.

Are Nuclear Waste Cells Really Aiming at 30‑Year Drone Endurance?

In mid‑2026, outlets like Defense One and TechRadar reported that lightweight nuclear‑waste batteries emerging from Rads to Watts could “fuel drones for 30 years via trickle charging,” sketching scenarios where small radiovoltaic cells slowly charge conventional batteries to keep drones aloft or operational for decades. Those reports associate Rads to Watts with nuclear waste isotopes and long‑endurance drone concepts, and social posts amplify the idea under labels like “Project SYMPHONEE” and unnamed team configurations. The appeal is obvious: if nuclear waste can be turned into compact power cells with multi‑decade lifetimes, the logistics model for aerial surveillance and remote sensing would change radically.

The primary DARPA documents, however, are more cautious and less specific. The Rads to Watts program page and solicitation discuss “high‑power nuclear radiation” sources but do not name particular isotopes such as strontium‑90, polonium, or carbon‑14, nor do they quantify “long‑lived” as 30 years. DARPA’s FAQ even emphasizes that proposer teams have “total flexibility” in choosing radiation sources, including the possibility of using reactors, while also warning that neutrons and gammas from such sources pose conversion and schedule challenges. Moreover, the solicitation explicitly states that unit cells are expected to use “small quantities of radioisotopes within licensing limits,” underscoring the regulatory boundary conditions around any high‑activity source. Put bluntly: the conceptual link to nuclear waste–powered drones is real in public discourse, but neither 30‑year lifetimes nor specific isotopes are yet grounded in publicly available DARPA documentation.

Physical Limits: Why Kilowatt Radiovoltaics Are So Hard

To understand why Rads to Watts is viewed skeptically by many scientists, it helps to revisit the physics of existing nuclear batteries. Popular coverage of carbon‑14 “diamond batteries” correctly notes that one gram of C‑14 yields on the order of tens of microwatts—far too little to power a smartphone without hundreds of kilograms of material, let alone a drone. Academic and government work on betavoltaic and alphavoltaic devices, including efforts at Lawrence Livermore National Laboratory, have focused on millimeter‑scale sources powering micro‑watt to milli‑watt loads for niche applications such as sensors, medical implants, and off‑grid IoT devices.

The limiting factors are not only decay power per gram, but also radiation‑induced damage in semiconductor materials. Alpha particles at several MeV can create hundreds of lattice vacancies in silicon with each impact, rapidly degrading device performance.[Prib’s Lab transcript] Wide‑band‑gap materials like synthetic diamond and silicon carbide are far more tolerant, but they are expensive or industrially demanding. Even with ideal materials, achieving kilowatt‑level electrical output from radioisotope decay within licensing limits requires either very high conversion efficiency, use of higher‑power isotopes, or architectural tricks such as concentrating radiation fields or stacking large arrays of cells. That is why mainstream analyses treat claims of “kilowatt nuclear coin cells” as aspirational and warn of a gap between hype and current engineering reality.

Regulation, Safety, and the Drone Use Case

Any discussion of nuclear‑waste‑powered drones must grapple with regulation and safety. The U.S. Nuclear Regulatory Commission tightly controls the use, transport, and containment of radioisotopes, and no nuclear batteries have yet been approved for unrestricted public deployment. Airborne platforms pose particular concerns: crash scenarios, dispersal of material, and adversary capture are all more complex than for stationary remote sensors. DARPA’s solicitation reflects this environment indirectly by stressing that unit cells must use only “small quantities” of isotopes, within licensing limits, and leaving procurement of sources to experienced teams.

Historically, such concerns have killed prior nuclear drone projects. A Sandia National Laboratories and Northrop Grumman effort to explore nuclear‑powered unmanned aircraft concluded with only a feasibility study and no hardware; the concept was deemed “not worth the risks involved” and shut down. That history shapes how both policymakers and engineers read today’s proposals: the potential operational advantages are enormous, but the political and safety costs are equally large. It is telling that Rads to Watts is framed around unattended, long‑lived power for “remote or power‑scarce environments” rather than explicitly advertising armed drones as flagship applications.

Hype, Fraud, and the Need for Verification

Rads to Watts is emerging in a landscape already crowded with nuclear battery claims. Startups such as NDB loudly promised diamond batteries that would power electric vehicles for decades and spacecraft indefinitely, only to face an SEC fraud lawsuit in 2023 for allegedly raising money without a working prototype. At the other end of the spectrum, companies like Infinity Power and Arkenlight pitch much more modest devices—tens of milliwatts over many years—for specialized scientific and industrial uses, carefully backed by lab data and peer‑reviewed work.

This history explains both the excitement and skepticism that greet DARPA’s kilowatt radiovoltaic vision. On one hand, DARPA’s track record in turning exotic physics into practical systems is strong, and the formal structure of Rads to Watts—the power contest, time‑capsule tests, and defined unit‑cell architecture—suggests a serious attempt to quantify progress. On the other, the concrete, independently validated data still sit firmly in the low‑power regime. Claims that small nuclear‑waste cells will soon power drones for 30 years are, at this stage, extrapolations from early contracts and program language, not results from peer‑reviewed tests or public performance reports.

DARPA plans 30-year endurance nuclear waste batteries to power next-gen drones, says report — project SYMPHONEE aims to harvest Strontium-90 to power persistent military drones https://t.co/d1CIzcZI2v

— Kelly (@mekellwhyy) July 5, 2026

What to Watch Next: Evidence That Will Clarify the Drone Question

For readers trying to discern whether nuclear waste will genuinely power next‑generation drones, the most important signals over the next few years will be technical, not rhetorical. Publication of detailed Rads to Watts solicitation documents, down‑select results from linear‑accelerator endurance tests, and any peer‑reviewed data on devices approaching the >10 W/kg target will reveal how far the field has moved beyond lab‑scale milliwatt batteries. Likewise, any NRC decisions on licensing radiovoltaic systems for airborne or mobile use will indicate whether regulators are willing to contemplate drone deployment rather than confining nuclear batteries to stationary or space applications.

In parallel, independent audits or FOIA‑released reports on Rads to Watts and related contracts—naming isotopes, lifetimes, and application envelopes—will either substantiate or temper current media narratives about 30‑year endurance and nuclear‑waste‑derived flight. Until that evidence emerges, the most defensible reading is that DARPA is pursuing a high‑risk, high‑reward research program aimed at pushing radiovoltaics into a new power regime; nuclear‑waste‑powered drones remain an intriguing, plausible application if the physics and engineering pan out, but not yet a proven or near‑term reality.