NEWS
Russia’s Starlink
Europeans are not the only ones looking to build sovereign alternatives to Starlink. Russia is reportedly spending billions on its unfinished Rassvet constellation. The latest filing, Rassvet-4, specifies 374 satellites operating at an altitude of ~800km. The programme is targeting 900 satellites by 2035.
In comparison, Starlink operates at an altitude of ~500km and already has ~10,000 active satellites, targeting 15,000 by 2035.
Satellite people will rightly point out that I’m comparing apples to oranges. A higher orbit means each satellite sees more of the ground, so Russia needs fewer of them. But it also means that Rassvet won’t deliver quite the same capability that Starlink does.
A constellation at 800km covers roughly twice the area per satellite, so Rassvet needs fewer satellites. But each link is weaker and capacity per km² is lower than at Starlink’s 500km altitude.
Translating that into battlefield capability, it means (roughly) that Russian soldiers using Rassvet will have to contend with bulkier, higher-powered terminals that are easier to detect and jam. And patchy coverage while the fleet is small. Whereas Ukrainian soldiers using Starlink benefit from small, cheap terminals and dense overlapping coverage that carries video from many drones at once. If Starlink is available to them.
By comparison, Europe’s IRIS² has first launches in 2029, initial services in 2030 and full operations in 2032, with 348 satellites. Eutelsat’s OneWeb already has ~650 satellites in orbit, but it was built as a commercial network. ICEYE and Nokia’s recently announced network for Ukraine plans first launches in 2028.
The implications for the war arrive slowly. According to Ukrainian analysts, the first Rassvet satellites already give two passes a day of more than an hour over Ukraine. In the medium term such a capability could convey an advantage to Russia, largely because it would gain a network it controls before Europe does.
In the short term, Ukraine still has access to Starlink and the battlefield advantage that comes with that, which is considerable. Starshield-enabled medium-range attack drones have helped Ukraine gain the upper hand in several areas of the front this year. But last month’s throttling of the service shows how conditional that access can be.
Shield AI buys Echodyne
I wrote last week about the acquisition of Rune by Salesforce, which felt like an early exit for Rune. And a disappointment for its investors. The same could not be said for Echodyne, a builder of compact radars for counter-drone systems, which was acquired by Shield AI for $5.3B.
Shield AI is known for Hivemind, its autonomy software, and the V-BAT drone (which has reportedly not performed well lately). Presumably it plans to wed Echodyne’s radar to its autonomy software. And pipe some of Echdyne’s revenues into its cashflows.
However perhaps the more interesting part is the involvement of Blackstone, which provided $1B of structured equity for the acquisition. Last month it spent $1B building an interceptor guidance technology. Clearly the sensing and guidance layer is where it thinks the value sits. I suspect we will see more of Blackstone in defence tech deals.
NATO Opens a Counter-Drone Marketplace
NATO launched its counter-drone marketplace, four months after announcing it. The NATO Communications and Information Agency selected 16 systems from five companies: Alta Ares, Axon, Northrop Grumman, OSL UK and Quantum Systems.
That means that NATO allies can now buy straight from the catalogue without running a further tender, and NATO has set aside €3.1M so they can trial systems before committing.
Interestingly, Franco-Ukrainian Alta Ares won six of the 16 lots, more than any other vendor. Largely on the basis of cost, it seems. And every system had to be ITAR-free, a requirement that still admitted Axon and Northrop Grumman (!).
FUNDRAISING
DEFENCE
Onodrim Industries, a Dutch company executing a roll-up of established businesses across aerospace and defence, raised a $300M round at a $1.3B valuation. Onodrim Industries is an Expeditions portfolio company.
Harbinger, an American developer of electric trucks, raised a $240M Series D co-led by Tiger Global, Capricorn’s Technology Impact Fund, and Ridgeline.
Corridian, an American developer of uncrewed underwater vehicles (UUVs), raised a $38M seed led by Point72.
Lumbra, an American developer of agentic operating systems for defence, raised a $28M Series A co-led by Point 72 Ventures and Ansa Ventures.
Primordial Labs, an American developer of natural-language software that lets military crews command sensors and other systems, raised an $18M Series A led by Ballistic Ventures.
Aformic, an American developer of robotics-for-manufacturing, raised a $6.95M seed led by Veteran Ventures Capital.
UADamage, a Ukrainian developer of defining technology, raised an undisclosed amount from Pravo Ventures and Ondas Capital.
SPACE
CesiumAstro, an American developer of communications systems, satellites, and space infrastructure, raiseda $630M Series D at a $4.5B valuation.
Reflect Orbital, an American builder of orbital mirrors to redirect sunlight for solar power and nighttime illumination, raised a $73.5M round.
Tumbleweed, an Austrian company conducting microgravity missions and develop a return-to-Earth vehicle, raised a €6.9M seed led by Pact.
Marble Imaging, a German provider of high-resolution satellite imagery, raised a €6M seed led by Cherry Ventures.
MuRay, a German developer of muon imaging technology for industrial components such as rocket engines, raised a €3.5M round led by First Momentum Ventures.
QUANTUM, CYBER & AI
Nominal, an American developer of technology for AI agents to analyze, test, and monitor complex systems, raised a $240M Series C at a $3B valuation, co-led by Lightspeed Venture Partners and Insight Partners.
Quint Computing, a German developer of superconducting quantum processors which integrate error correction into chip architecture, raised a €9.1M round led by Vsquared Ventures.
Physicl, a French developer of 3D environments and synthetic training data for physical AI models, raised a $10M seed led by Inovia Capital.