Early friction.
Lasting advantage.
Something is shifting in the funding landscape. Software is still dominant, but its share of startup capital has been sliding for years, while hardware is slowly gaining ground. It is not a dramatic turn, but it is enough to remind us that progress still depends on the physical world and on the teams willing to do the slow, unscalable work that real technology requires.
This week, we look at why that early, unscalable phase is not a setback but a useful part of building hardware, and how early friction often becomes the knowledge that later makes scale possible.
Nordic Hardtech Partners 🤝
- Lightbringer: AI-powered patent service helping tech companies protect their edge fast, with in-house legal expertise. Book an intro.
- Recuro: Growth partner for hardtech and deeptech. Scale with profit.
- SISP (Swedish Incubators & Science Parks): The network for 60+ incubators and science parks in Sweden. Find yours here.
- The Yard: Gothenburg’s hub for co-working, community, and serious hardware. Explore their memberships.
Nordic Hardtech Radar
This Week: Why scale is a reward, not a strategy.
The First Mile: Where Hardware Companies Are Actually Made
Hardware does not scale at first. It’s not meant to. Real progress begins when technology meets the physical world, and the early, unscalable work becomes a source of understanding rather than a delay. For many Nordic teams, this is where the moat forms.
Capital is slowly shifting in the same direction. Software still dominates, but its share of startup funding has fallen from 83 percent in 2016 to 64 percent in 2025, while hardware and deep tech are gaining ground. The shift does not make hardware easier, but it reinforces a basic truth: scale only works when the early groundwork is solid.
Early pilots, calibration work and low-fidelity prototypes are still unavoidable, and they remain the stage where teams learn the constraints they later build on. Several Nordic companies show how this unscalable first mile plays out in practice.
FleetMQ: learning the system before it automates
FleetMQ builds low-latency data streaming software for autonomous and remote-operated machines, where timing and precision are critical. Making the technology reliable has required direct validation in the field, through early pilots in road vehicles, boats and space robotics projects. Each environment forces the team to understand real-world communication before any scalable automation is possible. The company’s early prototypes were refined in WASP test arenas, giving the team a structured way to validate the system in real conditions.
Heimdall Power: installing on the real grid
Heimdall Power works in a domain where failure is not an option. Its sensors are installed on live power lines across Europe, in conditions no test lab can fully reproduce. Each installation adds to the team’s understanding of how grids behave in practice and strengthens the technology over time. In 2025, TIME named the company’s Neuron device one of the year’s best inventions, after early deployments in the United States increased transmission capacity by 63 percent at peak demand and delivered projected savings of around three million dollars.
Single Technologies: precision built in small steps
Single Technologies advances single-molecule imaging through slow, deliberate iteration. Calibration, repeatability and industrialization steps cannot be rushed, and precision builds over time. The company secured SEK 140 million in EU funding earlier this year, and an additional SEK 30 million from existing owners to move from research toward commercialization, a vote of confidence in the value of that long groundwork.
Skyfri: solar assets that behave differently everywhere
Skyfri operates in a part of the energy system where physical variation is constant. Each solar site behaves differently depending on weather, installation history and equipment quality, so early deployments require hands-on calibration and long observation periods. In 2025 the company reported that its technology revenues were on track to triple year over year, driven by growth across more than 150 solar sites in Europe and the United States, showing how software-heavy energy technology still depends on slow, real-world learning before scale can take hold.
And a 300-kilo reminder ...
Flatfrog offers a clear illustration of how early, unscalable work shapes long-term success. When Ola Wassvik co-founded the company in 2007, the first multitouch prototype was neither refined nor lightweight. As he told Nordic Hardtech earlier this year, “The first prototype weighed 300 kilos. It was built with consultants, duct tape, and a lot of optimism.” Its purpose was to expose real-world constraints before any refinement. As the company grew, engineering progress was paired with an aggressive IP strategy. According to Ola, “One cannot out-price China, our only shot was to out-innovate, and protect every move.”
Scale as a reward, not a strategy
The broader lesson is clear: real scale follows understanding, not the other way around. Hardware companies win by learning the physical reality of what they are building long before growth is possible. That knowledge comes from slow pilots, site visits, crude prototypes and early friction that others never encounter. It’s not fast or glamorous, but it anchors everything that comes later.
Lately ⏮️
Selected news from the hardtech ecosystem
- Vidde’s electric snowmobiles just claimed the 2025 E-Prize. A win built on years of stubborn engineering, and a sign that Sweden’s circular EV niche is getting real.
- Nordic Innovation launches “Nordic Circular Investments” to map what actually makes circular business models pay off. A finance-first push to show which cases scale — and why.
- AlixLabs just raised €14.1 million to scale its APS process for atomic-level patterning. A simpler, low-energy alternative to EUV, and a potential shift in how next-gen chips get made.
- Arkeon claims a fix for one of quantum’s biggest bottlenecks: tuning qubits after fabrication. A Chalmers-validated method that could cut waste and make large-scale chips viable.
- Finland’s Donut Lab just opened its DonutOS platform, giving small mobility teams a way to design and validate hub-motor vehicles without hardware in hand.
Up next ⏭️
What's brewing in the community?
- Tonight in Stockholm! Founders Live returns with five 99-second pitches and live voting. A quick way to meet the city’s early-stage founders — doors open at 17:00.
- Nominations for the Nordic Scaleup Awards are now open. A reminder that scale is more than speed, the jury looks for sustainable, well-built growth. Deadline 3 March.
- Arkus AI hosts a hands-on session on health-focused AI agents, with robotics insights from Uppsala University. A practical look at where medtech tools are heading next. December 2nd in Stockholm.
Nordic Hardtech Funding List
See who’s raising, scaling and building across the Nordics.
Nordic Hardtech is a platform for the hardtech ecosystem — sharing knowledge, bringing inspiration, and building community through podcasts, newsletters, events, and more.
💸 Fuel up. Scaling takes capital — The Nordic Hardtech Funding List shows you where to find it.
💥Get involved. Pitch a topic. Join the podcast. Get on the investor list– if you're hardtech.
→ Spotify | LinkedIn | nordichardtech.com
📩 hello@nordichardtech.com
