Estonia's Golbriak raises 4 million euros to mass-produce satellite laser links

Golbriak Space, an Estonian company that designs and builds laser communication terminals for satellites and aircraft, has raised 4 million euros to industrialise them. The seed round was co-led by Join Capital and PhotonVentures, with the aerospace accelerator Takeoff taking part. The investment went into Golbriak's Italian holding company, Iribits S.r.l., and will fund the Estonian operating subsidiary, Golbriak Space OÜ, as it scales production.

The size of the round is not the interesting part, because 4 million euros is small even by Estonian standards. The problem the company says it has solved is. Laser links between spacecraft have been demonstrated for years and their advantages are understood. What has held them back is manufacturing: terminals that can be built by the thousand, mounted without a team of engineers, and pointed at a partner moving at kilometres per second.

Why laser beats radio, and where the difficulty sits

Radio is the default for satellite communication and it is running out of room. The spectrum is congested, and using it is licensed, so operators wait for coordination and pay for the privilege. Light needs neither.

• Optical beams diverge three to four times less than radio beams at comparable frequencies, according to the company, which means less of the signal is lost on the way to the receiver.
• A narrow beam is harder to intercept or jam. An adversary would need access to the beam path or precise alignment on the receiver, rather than an antenna and the right frequency.
• Optical communication is not regulated the way radio is, so there is no international frequency coordination and no licence fee before a link can be switched on.
• The same narrowness that makes a laser private makes it fragile. The two ends have to find each other and stay aligned, and a satellite does not hold still.

That last point is where the engineering effort has gone, and it is what Golbriak claims to have reduced: its terminals are built to depend less on the pointing accuracy of whatever carries them, which makes them easier to mount in multiples on one spacecraft or aircraft. The company designs the optical, electronic, mechanical and software parts itself rather than assembling bought-in modules, and sells two product families on a single software-defined architecture.

What is actually in the box

• ERMES, for space: from 2U in size, where one U is a ten centimetre cube and the standard building block for small satellites, from 1.2 kilograms, 15 watts of peak power, 2.5 gigabits per second in low Earth orbit and link distances from 10,000 kilometres on the base model.
• ARES, for aircraft: two 2U units, 1.8 kilograms, 15 watts of peak power, and 2.5 gigabits per second or more through the atmosphere.
• A downlink service sold by the minute, listed from 18 euros per minute at 1 gigabit per second.

Note the gap between the front page of the company's site, which promises hundreds of megabits per second, and the product pages, which quote 2.5 gigabits per second. Both are the company's own figures, neither has been independently verified, and a buyer with a real link would test them.

The demand case is arithmetic about constellations

The commercial argument is straightforward. Satellite networks are getting larger, more interconnected and heavier with data, and a single constellation programme can need hundreds or thousands of terminals. Golbriak says it is in advanced negotiations on programmes involving several thousand systems, with customer-funded deployments starting next year and further units being prepared for airborne use.

The claim that carries the weight is the one about paying customers. It is the line that separates a funded laboratory from a supplier, and it is also the hardest to check from outside. PhotonVentures, which co-led the round, states its own conclusion about where that leads.

What the announcement does not say

• No customer is named and no contract value is given, for negotiations that only matter if the several thousand figure converts into orders.
• No revenue is published, so there is no way to judge how many terminals anyone has actually bought.
• The 4 million euros is an industrialisation round rather than a scaling one. A production line for hardware that has to survive launch vibration, vacuum and radiation is not built cheaply, which suggests the next round decides how large this company becomes.
• The company is an Italian holding with an Estonian operating subsidiary, a structure that will matter if it seeks public support or defence procurement in either country.

Golbriak is not in our directory, so nothing here rests on a directory profile. The technical detail comes from the company's own pages and the funding detail from Tech.eu.

What to watch

• Whether the negotiations convert, and who the customers are. Several thousand terminals is a number that either becomes a contract or quietly stops being repeated.
• Whether anyone independent tests the terminals. Inter-satellite laser links have a short list of proven suppliers, and a European entrant on that list would be the story.
• Whether the airborne product ships. ARES is the half of the portfolio with no constellation behind it and the clearest route to defence customers, including in the Baltics.
• Whether European constellation programmes buy European terminals. Institutional demand is the difference between a supplier and a component maker with good technology.
• Whether the downlink service becomes a real product line. Selling minutes of capacity is a different business from selling hardware, with different margins and different customers.
• Whether the next round is larger. Capacity for space hardware is bought with money, and 4 million euros does not buy much of it.

Optical terminals are a market where the physics was settled decades ago and the manufacturing was not, which is an odd place to compete and an odd place for a small country to win. Golbriak's case is that it has working products, paying customers and demand at a scale its competitors cannot serve quickly. This round is enough to test that case and not enough to prove it. The proof arrives as terminals in orbit, and the company says that starts next year.