Pluvio · Vancouver, Canada

Liquid droplet radiators for high-power spacecraft.

Every watt a spacecraft uses becomes heat it has to radiate into space. Pluvio is developing a radiator that does this with a sheet of recirculating droplets instead of solid panels, aiming to reject the same heat at a fraction of the mass.

Early stage · Building our first bench prototype

How it works Work with us Read the thesis

Illustration of a droplet sheet. Not Pluvio's design; not to scale.

01 · The constraint

In space, the only way to get rid of heat is to radiate it.

On Earth, data centres move heat away with air and water. In orbit there is nothing to carry it. A spacecraft sheds waste heat by radiating it from hot surfaces into cold space, and how much heat a surface can shed is set by its temperature. Electronics cap that temperature, so the only way to reject more heat is more radiator area, growing in step with every kilowatt added.

Fig. 1: Idealised double-sided radiator at 325 K, emissivity 0.9, no environmental heat load. Squares and court drawn to the same scale. Real radiators need more area than this.

At 1 MW, that is roughly 878 m² of radiator under ideal assumptions. At a gigawatt it is about 0.88 km², approaching a square kilometre.

02 · Why now

Orbital compute is being funded. Heat rejection is what's left.

Orbital data centres have moved from concept to capital. Starcloud raised $250M at a $2.3B valuation in August 2026, bringing its Series A to $420M.¹ SpaceX has applied to the FCC for a constellation of up to one million orbital data centre satellites.² In October 2026 Google launched the first prototype satellite for Project Suncatcher, its research effort to run AI compute in orbit.³

Four things set the cost of computing in orbit: launch, solar power, compute density and heat rejection. The first three ride industrial learning curves: rockets are getting cheaper, solar cells lighter, chips more efficient per watt. Heat rejection is bounded by physics and improves slowly.

Launch price$2,700/kg$300/kg (model base case)$100/kgFree launch

Data table

Launch price	Earth	Orbit: launch	Orbit: solar	Orbit: heat rejection	Orbit total	Heat rejection share
$2,700/kg	$1.63B	$2.84B	$1.21B	$7.25B	$11.30B	64%
$300/kg (model base case)	$1.63B	$0.32B	$1.21B	$3.84B	$5.37B	72%
$100/kg	$1.63B	$0.10B	$1.21B	$3.56B	$4.88B	73%
Free launch	$1.63B	$0.00B	$1.21B	$3.41B	$4.63B	74%

Fig. 2: Cost of power and cooling for a gigawatt of compute, per year, on Earth and in orbit, in our model (Pluvio cost model v3, 1 October 2026). Billions of dollars per GW-year. Compute hardware is excluded because it costs the same in both places. Heat rejection is radiator hardware plus the launch of its mass. Model output, not measured data.

Cheaper launch makes the radiator a bigger share of the bill.

In our model, heat rejection is nearly two thirds of the cost of power and cooling in orbit at today's launch prices, nearly three quarters at Starship's target price, and about three quarters even if launch were free. Most of the radiator's cost is the hardware itself, so cheaper launch barely dents it: every dollar taken out of launch leaves the radiator as a larger share of what remains.

The same limit applies to any spacecraft that has to reject more heat than solid panels can carry economically. Orbital compute is the first market to hit it at scale.

03 · Our approach

A radiator made of droplets.

A conventional radiator pumps coolant through a large solid panel, and the panel does the radiating. A liquid droplet radiator removes the panel. The hot coolant is sent across open space as a sheet of fine droplets. Small droplets have an enormous surface area for their mass, so they radiate heat efficiently as they travel. On the far side, a collector recaptures the cooled droplets and returns them to the loop.

Fig. 3: Schematic of a generic liquid droplet radiator. Not Pluvio's design; not to scale.

The generator forms a sheet of hot droplets.
The droplets radiate heat to space in flight.
The collector recaptures the cooled droplets.
A pump returns the fluid through a heat exchanger to pick up heat from the payload again.

Mass.

In our modelling, a droplet radiator is 52% to 74% lighter per kilowatt than a two-sided solid radiator, under assumptions that favour the panel. Assumptions

Resilience.

The droplet sheet is relatively immune to micrometeoroids; only the generator, collector and piping need armour.

Turndown.

When the payload's heat load drops, a droplet radiator can be turned down or switched off. A solid panel keeps radiating regardless, so in cold phases its coolant often has to be kept warm with heaters to stop it freezing.

Stowage.

Because the radiating surface is a sheet of droplets that packs down to a small volume of liquid, an LDR promises compact stowage: there are no large panels to fold, only the generator, collector and the structure between them.

The advantage grows with power.

A panel radiator scales with area, and area scales with power: under ideal assumptions a megawatt needs about 878 m² of panel and a gigawatt about 0.88 km² (Fig. 1). Every square metre has to be launched, folded, deployed and protected, and a bigger panel is a bigger target for micrometeoroids. A droplet radiator still needs more fluid and a wider sheet as power grows, but its radiating surface is formed in orbit from a small volume of liquid. In our modelling it is 52% to 74% lighter per kilowatt, so the mass it saves grows with every kilowatt added. That is why we are building for spacecraft in the 100 kW to megawatt class.

04 · What has to be true

The hard parts, and how we plan to test them.

A droplet radiator is simple to describe and difficult to build. These are the questions that decide whether it works. Each one maps to a stage of our test programme.

Droplet generation. Can we produce a uniform, well-aimed sheet of fine droplets, reliably and for long durations?
Bench
Recapture. What fraction of droplets do we recover, and does fluid loss over a mission lifetime stay within budget?
Bench · Vacuum
Vacuum behaviour. How does the fluid behave, including evaporative loss, across the operating temperature range in vacuum?
Vacuum
Microgravity. Does the sheet stay stable, and does collection still work, without gravity and while the spacecraft manoeuvres?
Microgravity
Contamination. Can we keep fluid off optics, solar arrays and other sensitive surfaces?
Vacuum · Orbit
Integration. Can the droplet loop couple cleanly to a payload's thermal system and operate for years?
Flight

05 · Where we are

Where we are.

Pluvio started in 2026. Our radiator currently exists as a design and a model, and we are building our first bench prototype now.

BUILDING
Bench
Generate and recapture droplets in the lab; measure recapture fraction and heat rejected.
NEXT
Vacuum chamber
Repeat under vacuum across the operating temperature range.
PLANNED
Microgravity
Test sheet formation and collection on parabolic flights.
PLANNED
Flight demonstration
An in-orbit demonstration with a design partner.

We update this page as each stage completes.

06 · Direction

The thermal system for high-power spacecraft.

We don't plan to build satellites or run data centres. We want to be the independent supplier of heat rejection to the companies that do: a subsystem an operator can specify, integrate and buy, backed by flight data showing it works. Spacecraft operators favour hardware that has already flown, so the first droplet radiator qualified in flight will have a head start. We intend to build that radiator.

Beyond orbital compute.

Orbital compute is the first market to hit the heat limit, but not the only one. We see two more places where a droplet radiator could matter:

Lunar bases. A permanent base on the Moon will run power systems that have to shed their waste heat with no air or water to carry it, and every kilogram of radiator has to be landed. A radiating surface formed from a small volume of liquid could cut that load. Lunar gravity and dust change the design problem, and our current test programme does not cover them yet.
Manufacturing in orbit. As in-space manufacturing moves from small experiments to production, furnaces and processing lines will produce waste heat on a scale that today's satellites have never had to reject. The same argument applies there: the more heat, the more a lighter radiator saves.

07 · Team

Who we are.

William Semmens
Co-founder and CEO
Trystan Semmens
Co-founder and CTO

08 · Work with us

Work with us.

Building a high-power spacecraft? We're looking for design partners for our first flight demonstration.
Investing in space infrastructure? Ask for our technical brief and cost model.
Engineers. We want to hear from thermal-fluids and spacecraft engineers who find this problem interesting.
Government and research. We'd like to work with agencies, labs and universities on testing and development.

Emailwilliam@pluvio.space