Reflect Orbital wants to sell sunlight after dark. On October 14 the Los Angeles startup bought two dedicated launches from Skyroot Aerospace, adding working rockets to a plan that until this week lived mostly in pitch decks and a single regulatory filing.
The FCC licensed one demonstration satellite in July. The 50,000-mirror constellation behind it is still unapproved and contested.
Getting to orbit is the easy part. Delivering controllable light on demand, without disturbing the night sky, is the hard one.
The deal is small in dollars and large in signal. Skyroot did not disclose the price of either mission, and Reflect Orbital did not say when they fly. Booking two rockets anyway is the first hard commitment in a business built on mirrors that do not exist yet.
Reflect Orbital's 2035 constellation target
From two demonstration satellites in 2026 to more than 50,000 by 2035, on the company's own published roadmap. · Reflect Orbital, 2026
Two rockets, two versions, no dates
The contract covers one flight on Vikram-1 and one on Vikram-1U. Vikram-1 reached orbit for the first time on July 18, 2026, the first privately developed Indian rocket to do so, and carries up to 350 kg to low Earth orbit.
Vikram-1U has not flown. Skyroot expects its first mission in the first quarter of 2027, with additional first-stage boosters lifting capacity to roughly 550 kg. Booking both versions spreads the risk across a proven vehicle and an unproven upgrade of the same design.
Reflect Orbital is the only payload on both flights, a spokesperson told Payload. A dedicated launch buys schedule control and orbital choice that a rideshare slot cannot offer. It also means paying for the whole rocket.
What the company is actually selling
Reflect Orbital's product is redirected sunlight, not generated power. A spacecraft that can see both the Sun and a target area tilts a large reflector so daylight lands inside a spot about 5 km wide. Rotating the mirror, or letting the geometry slip, switches the light off.
The published roadmap is unusually specific for a company this young. In 2026 it promises 0.1 lux for five minutes, roughly full moonlight. By 2028 the target is up to 100 lux for two hours, comparable to an indoor work area. By 2035 it projects up to 36,000 lux, close to daylight, from a constellation of more than 50,000 satellites.
| Year | Satellites | Lighting service | Energy service |
|---|---|---|---|
| 2026 | 2 | 0.1 lux / 5 min (full moon) | Testing |
| 2028 | 1,000+ | Up to 100 lux / 2 h | Testing |
| 2030 | 5,000+ | Up to 5,000 lux / minutes | +1% capacity factor |
| 2035 | 50,000+ | Up to 36,000 lux / hours | +20% capacity factor |
Source: Reflect Orbital published roadmap, 2026
The energy product is separate and far more modest. Reflect Orbital estimates its mirrors could add 1% to a solar farm's capacity factor by 2030, and 20% by 2035. Those gains assume a typically overbuilt array. For Southern California the company projects 0.75%; for Germany, 1.4%.
The physics of a 5 km patch of light
A reflector in low Earth orbit crosses the sky in minutes. Reflect Orbital needs the mirror, the Sun and the ground target to line up at the same time, which shortens each pass and constrains the orbit the spacecraft can use.
Brightness scales with the collecting area of the reflectors and with their distance from the ground. One small satellite under a favourable sun angle delivers a faint fraction of moonlight; the company's own 2026 target is 0.1 lux for five minutes. Reaching street-lighting levels takes many satellites pointing at the same spot at once.
That is why the roadmap grows so steeply. Two satellites in 2026 become 1,000 by 2028 and 50,000 by 2035. Duration and intensity both come from overlap, not from one bigger mirror.
Weather and orbital geometry eat into usable hours. A satellite in low Earth orbit spends a large share of each pass in Earth's shadow, and clouds block the beam however carefully the orbit is planned. Reflect Orbital sidesteps some of this by selling scheduled, locally approved windows instead of continuous light.
The hard part is coordination. Thousands of independently moving mirrors have to hold one target long enough for the light to be worth paying for.
The regulator said yes. Two thousand comments said no.
In July 2026 the FCC approved Earendil-1, Reflect Orbital's first demonstration spacecraft. The application drew more than 2,000 public comments, many of them opposed.
Environmental groups asked the commission to cancel the licence. They argue that bright beams switched on at night could disturb wildlife, interfere with astronomical observation, and disrupt human circadian rhythms. The American Astronomical Society urged its members to take part in the review.
We will continue to work with astronomers, local communities and regulators to earn the right to scale.— Ben Nowack, co-founder and CEO, Reflect Orbital
The idea is not new. In 1993 Russian engineers unfolded a 65-foot mirror in orbit and steered a beam of sunlight across Europe. The Znamya experiment worked; the funding did not. What changed since is the cost of small satellites and the appetite of private capital to fund a demonstration before a market exists.
Why an Indian rocket
Reflect Orbital's binding constraint is launch, not light. SpaceX has stopped taking new reservations on its Transporter rideshare programme after 2029, and operators have started buying whole vehicles to guarantee slots. Portal Space Systems took a full Falcon 9 in August. Maverick Space Systems then filled the spare capacity with 10 Kepler satellites.
Skyroot gives Reflect Orbital a second supply line outside the United States. The Hyderabad company became India's first space-tech unicorn in May 2026, raising nearly $60 million at a $1.1 billion valuation, and its Vikram-1 launch in July was the country's first private orbital flight.
The economics have to catch up
The lighting business raises fewer questions than the energy one. Search-and-rescue teams, remote industrial sites and event organisers can pay for a scheduled patch of light without a grid connection, and the company says it will only serve locations approved by local authorities.
Solar augmentation is harder to sell. A 1% capacity-factor gain by 2030 will not move a utility's investment case. As we wrote in our October piece on space-based solar, the first paying customer for power from orbit arrived before the costs did. Reflect Orbital's own projections put the meaningful energy number in 2035, a decade away.
What would have to be true by 2028?
Probability: 40% — the hardware is simple next to the regulatory and business questions, and neither the upgraded rocket nor the market is proven.
🟢 Optimistic scenario (25%)
Consequences: Skyroot signs more Western customers; orbital illumination becomes a recognised category.
🟡 Base scenario (45%)
Consequences: The company survives on private capital; the energy service slides past 2030.
🔴 Pessimistic scenario (30%)
Consequences: Reflect Orbital narrows its plan and starts licensing its mirror hardware instead.
The first Vikram-1U flight date, currently Q1 2027.
Any FCC order wider than the Earendil-1 demonstration licence.
Reflect Orbital's first named paying customer for lighting or energy.
Whether Skyroot's manifest converts into repeat international orders.