BSEIF’s Venture division has published Heat Is the New Constraint in Orbit: Who Will Build the Radiators of the Space Economy, the seventh piece of official divisional research released by the association and the division’s first full paper after its Initial Statement. It was written by Tommaso Castelli Rivolta, Chris Vittorio Joannes Agnelli, Ahmet Selim Birışık, Haya Kret and Riccardo Maggioni, reviewed by Riccardo Gilli, Head of Venture, and Gregorio Pesente, and approved by Riccardo Gilli. It asks who will build the hardware that removes heat from spacecraft, and where in that chain early-stage investors could find value.
The paper starts from the physics. In a vacuum a spacecraft can shed heat only by radiating it, and the main cooling system of the International Space Station needs about 475 square metres and almost 7 tonnes of radiators to reject 70 kW. At that performance, by the authors’ estimate, one megawatt would take about 6,800 square metres and nearly 100 tonnes, some $145 million just to launch. Demand is rising on two fronts at once. There were 16,913 active satellites on 30 September 2026, of which 11,143 were Starlink, and each carries more power: Europe’s first Eurostar telecom satellites had at most 1–2 kW in total, while Airbus’s Eurostar Neo devotes 25 kW to its telecommunications equipment alone. Orbital data centres, from Starcloud’s H100 in orbit to Google’s Project Suncatcher, would be the most extreme case.
The analysis then maps the thermal value chain in five segments: radiators, heat pipes and loop heat pipes, materials, coatings and active systems. Deployable radiators, adaptive coatings and components of active cooling loops are the most open to newcomers. Heat pipes are protected by formal qualification standards, and heat pumps, which would let a smaller radiator shed the same heat, still have no commercial entrant. Capital is not following the constraint. Global investment in space ventures reached €11.7 billion in 2025, with €4.7 billion going to launch companies and €3.1 billion to satellite manufacturing, and Starcloud alone raised $420 million in 2026. The four thermal start-ups the paper profiles have disclosed about $25 million between them.
Those four came out of a screen of 26 thermal-hardware suppliers against four criteria, from funding stage to verifiability through public sources: SmartIR in the United Kingdom, with adaptive-emissivity coatings; Sophia Space and Beyond Reach Labs in the United States, with integrated passive radiators and deployable structures; and In Quattro in Italy, an ENEA spin-off building two-phase cooling loops, whose figures come partly from an interview with its chief executive. On exits, 86% of space start-up exits between 2000 and 2020 were acquisitions, and thermal deals so far are small: $17 million for Roccor and €19 million for IberEspacio. The large values sit with companies that reached markets beyond space, such as Boyd Thermal, acquired by Eaton for $9.5 billion.
The paper names four risks: that orbital data centres remain hype, that qualification cycles delay revenue, that customers are few and concentrated, and that primes build thermal systems in-house. Its investment thesis rests on three principles. Early-stage capital should back components that fit many platforms, position for space-for-space activities, where the constraint binds hardest, and use markets on the ground to de-risk the space case without replacing it, above all liquid cooling for data centres, expected to grow from $4.1 billion in 2026 to $27.7 billion by 2033. The report is published for information and education only and is not investment advice.
The full paper is available under Insights, in the Venture section. It can be read on the website, opened separately or downloaded as a PDF.