Eight hundred watts at 8.6 kilometers: what DARPA actually demonstrated

What the work claims

DARPA claims that the POWER Receiver Array Demo (PRAD) set a distance record for optical power beaming by delivering more than 800 watts of power during a 30-second transmission from a laser 8.6 km away.¹ Over the full test campaign, more than one megajoule of energy was transferred. The previous record was 230 watts of average power at 1.7 km for 25 seconds. DARPA states that the demonstration "obliterated all previously reported optical power beaming demonstrations for power and distance."

The receiver was designed by Teravec Technologies, led by principal investigator Raymond Hoheisel, with support from Packet Digital and the Rochester Institute of Technology. The test was conducted at the High Energy Laser Systems Test Facility (HELSTF) at the US Army's White Sands Missile Range, with participation from the US Naval Research Laboratory.¹

How it works

The PRAD receiver uses a compact aperture design that minimizes light escape once the laser beam has entered. Inside the receiver, the laser strikes a parabolic mirror that reflects the beam onto dozens of photovoltaic cells arranged around the inside of the device. These cells convert the optical energy back to electrical power.¹ The design is essentially a photovoltaic receiver optimized for monochromatic laser light rather than broadband solar radiation, which allows higher conversion efficiency at the specific laser wavelength.

The test geometry is critical. Both the transmitter and receiver were on the ground, meaning the laser beam traveled horizontally through the thickest part of the atmosphere. DARPA program manager Paul Jaffe noted that it is easier to send a power beam directly up or down relative to the ground because there is less atmosphere to traverse, and that PRAD was intentionally tested under maximum atmospheric impact.¹ This is the hardest case for atmospheric attenuation, not the easiest.

The POWER program's broader vision is an airborne optical energy relay network. The program seeks to design and demonstrate relays that can receive a ground-sourced laser beam, relay it through multiple airborne nodes, and deliver it back to a ground receiver. The PRAD test validated the receiver technology that would be needed at the endpoints of such a network. POWER Phase 2, with an Industry Day on May 29, 2025, is moving toward demonstrating integrated relays and vertical power transmission.¹

The strongest case

The strongest case is the raw improvement over prior work. The previous record was 230 watts at 1.7 km. PRAD delivered more than 800 watts at 8.6 km, which is a factor of 3.5 in power and 5.1 in distance, simultaneously. The receiver was designed and built in approximately three months, which suggests the approach is not a one-off laboratory artifact but a design that can be iterated rapidly. DARPA states the technology is scalable to higher power levels and can be integrated into platforms like unmanned aerial vehicles.

The 20 percent end-to-end efficiency figure, measured at shorter distances, is also significant. It represents the ratio of optical power out of the laser to electrical power out of the receiver. For a ground-to-ground horizontal path through maximum atmosphere, 20 percent is a credible result that suggests the receiver design is not the bottleneck. If the path were vertical (space-to-ground or ground-to-space), atmospheric losses would be lower, and the efficiency could be higher.

Where a skeptic should push

The efficiency measurement boundary matters. DARPA states that "more than 20 percent efficiency" was measured "from the optical power out of the laser to the electrical power out of the receiver at shorter distances." That is not the efficiency at the 8.6 km record distance. At 8.6 km through horizontal atmosphere, the efficiency would be lower due to atmospheric scattering and absorption. DARPA explicitly states that "efficiency was not the focus of this demonstration," and trade-offs were made to accelerate the design and build of the test receiver.¹

The 800 watts was delivered for 30 seconds. That is a demonstration duration, not an operational duration. A space power beaming system would need to operate continuously for hours, days, or indefinitely. Thermal management of the receiver at sustained power levels is a different engineering challenge than a 30-second burst. The receiver's photovoltaic cells absorb laser energy and convert a fraction to electricity, with the rest becoming heat. At 800 watts input and 20 percent efficiency, approximately 640 watts of heat must be rejected from the receiver during operation. Sustained operation would require active cooling.

The test is ground-to-ground. The POWER program's ultimate vision involves airborne relays and vertical transmission, neither of which was demonstrated in PRAD. The atmospheric path for a space-to-ground link is fundamentally different from a horizontal ground path: the vertical atmosphere is thinner, but the total distance is orders of magnitude larger (400 km for LEO versus 8.6 km for PRAD). Beam divergence over 400 km would require a much larger receiver aperture or a much tighter transmit beam than PRAD tested.

Finally, the POWER program page states the program "is now complete" and the content is available for reference purposes. The Phase 2 Industry Day was scheduled for May 29, 2025, but the program's current status and funding trajectory are not addressed in the news release.

What it means for power in space

The non-obvious implication is that the receiver design, not the transmitter, was the limiting technology. DARPA's statement that the PRAD receiver "broke through misconceptions about the limits of power beaming technology" points to the parabolic-mirror-plus-photovoltaic-cell architecture as the innovation. For space power beaming, this means the ground receiver for a space-to-ground link may be more tractable than previously assumed, because the PRAD design can capture a high fraction of incident light with a compact aperture. The specific mechanism is the internal reflection geometry: once light enters the receiver, very little escapes, which is the design choice that enabled the distance record.

The genuine threat is thermal. At 800 watts input with 20 percent conversion, the receiver must dissipate hundreds of watts of waste heat. In the PRAD test, this was manageable for 30 seconds in a ground environment with convective cooling. In space, a receiver on a spacecraft or a relay node has no convective cooling and must radiate all waste heat. The thermal design of a space-based receiver would be the dominant mass and complexity driver, not the photovoltaic conversion efficiency.

The opportunity is that the PRAD receiver technology is scalable and platform-integrable. DARPA states it can be integrated into unmanned aerial vehicles. If the same receiver architecture were scaled to the kilowatt or tens-of-kilowatts level needed for spacecraft power resupply, the design could serve as the receiving end of a space-to-space power link. The parabolic mirror geometry is wavelength-agnostic within the photovoltaic cell's responsive band, so it could work with different laser wavelengths optimized for atmospheric transmission.

The bottom line

What is demonstrated is a ground-to-ground optical power beaming record: more than 800 watts over 8.6 km for 30 seconds, with a receiver design that is compact, rapidly built, and scalable. What is not demonstrated is sustained operation, space-to-ground transmission, or system-level efficiency at the record distance. The 20 percent end-to-end efficiency is measured at shorter range, not at 8.6 km. Confidence is high that the PRAD receiver architecture represents a genuine advance in power beaming receiver technology. Confidence is lower that this specific test geometry translates directly to a space power beaming system, because the distances, durations, and thermal environments are qualitatively different. The reading would be strengthened by published efficiency data at the record distance and by a sustained-operation test of hours rather than seconds. It would be undercut if the receiver's thermal management proves unscalable above the kilowatt level.