Tusk IC's €15M Bet: Putting Satellite Antennas on Cheap CMOS

Fifteen million euros. That is the size of the Series A Tusk IC disclosed on September 17 — and its investors say it is enough to move the costliest part of a satellite internet terminal onto the same commodity silicon that already powers a car's radar.

Tusk IC designs millimetre-wave integrated circuits (ICs) in Antwerp. KU Leuven spun the company out in 2018, and for most of its life it sold engineering work and high-frequency measurement services. The satellite business is the part it now wants to industrialise.

A small round that stands out in Flanders

For a Belgian hardware company, the round runs large. Dealroom, the Amsterdam-based data provider, ranks it above the 90th percentile for comparable Series A deals in the country and values it at $17.2 million. The leads are specialists: Matterwave Ventures of Munich, the Dutch deep-tech fund FORWARD.one, and the Flanders Future Tech Fund, managed by PMV, the Flemish regional investment company. The Schaubroeck family joined them, alongside existing backers.

Vincent Hebbelynck, who heads technology venture capital at PMV, tied the deal to place rather than product.

Flanders has built a strong ecosystem at the intersection of semiconductors and space technology.— Vincent Hebbelynck, Head of Tech VC, PMV

The company spent eight years building millimetre-wave expertise inside a region that hosts both a chip cluster and a space industry. A €15 million round meant to connect the two reads less like a venture novelty than like industrial policy with a return target attached.

The half of satellite internet nobody prices

Satellite broadband has two halves. The first is the constellation — Starlink, Amazon's Leo, Europe's publicly funded IRIS². The second is the ground segment: the flat-panel antenna on a roof, a plane or a ship, pointing an electronic beam at a satellite crossing the sky. Ka-band, the high-frequency radio band operators use for high-capacity links, makes that beam harder to hold and the electronics behind it more valuable.

Electronic steering needs no moving parts. It needs an array of small antenna elements, each fed by a radio-frequency chip that shifts phase and controls amplitude. That chip, the beamformer, concentrates a terminal's semiconductor value. Builders have traditionally made it on gallium arsenide or specialty silicon germanium, both harder and pricier to manufacture than the mainstream bulk CMOS used for ordinary logic.

Its claim is that the beamformer can be built in that mainstream CMOS. If it holds, it changes cost and supply at once: bulk CMOS runs on high-volume lines that exist in dozens of fabs, not a handful of specialty processes. The company also points to record-low energy consumption — a real constraint, because a terminal's heat budget tracks how much power its electronics burn.

A chip with a European mandate

The commercial turn is recent. On July 6, Tusk IC announced an industrial contract from the European Space Agency (ESA), backed by BELSPO, Belgium's federal science policy office. It sits inside ARTES, ESA's Advanced Research in Telecommunications Systems programme, under a line the agency calls Industrial Competitiveness.

The mandate is narrow and unglamorous: take the ConnectKa chips and ConnecTile modules through design optimisation, qualification testing, pilot production and the start of volume manufacturing. ESA is funding the middle of the pipeline — the stretch between a working prototype and a part a customer can order in quantity.

By supporting Tusk IC's pioneering solution with record-low energy consumption and fabricated in mass-volume chip technology, ESA is demonstrating its commitment to strengthening Europe's competitiveness in satellite communications.— Constantin Siriteanu, Telecommunication Systems Engineer, ESA

The subtext is sovereignty. Europe is funding IRIS², its own secure constellation, while its terminal makers still lean on specialty chips and component suppliers concentrated outside the bloc. Moving the beamformer onto commodity silicon keeps one more layer of that stack inside Europe — and, in principle, cheaper.

Who buys the chip, and who else builds it

The buyer rarely sits at the consumer end. The company sells to the companies that build flat-panel antennas and satellite user terminals, which sell on to operators and integrators. Its modules serve low, medium and geostationary orbits — the three bands of altitude that different services occupy.

Three rivals matter. Sivers Semiconductors, listed in Stockholm, sells Ka-band beamforming chipsets. EnSilica, based in Britain, has taken ESA-backed orders for a digital beamformer aimed at broadband terminals. All three chase the same upgrade cycle: as low and medium Earth orbit constellations enter commercial service, the terminals that connect to them scale with the fleet.

What the investor is actually buying

The headline figure matters less than what it funds. A €15 million round buys a transition from prototype to volume production at a company that has already cleared an ESA qualification path. That sequence de-risks the technology, and hardware investors pay for exactly that.

The risks look just as concrete. The business runs fabless: it designs chips and pays a foundry to build them, so margins hinge on how much cost it can design out and how much pricing power it holds against Sivers and EnSilica. Pilot production is where deep-tech firms stall — yields, qualification and customer test cycles drain cash before revenue lands. The terminal market also tracks constellation capex, which carries its own history of delays.

As we wrote in September, the bottleneck in the space economy has moved from orbit to the ground. Satellites are launching; the terminals that make them worth launching are the constraint. A €15 million Series A aimed at the most expensive chip in that terminal is a small round attached to a very large question.

Will standard CMOS become the default beamformer process by 2029?

Development scenarios