Most weather forecasts you rely on are built on a surprisingly thin scaffolding of data. Satellites overhead, a scatter of ground stations, the occasional weather balloon. Between those points sits a vast, under-measured atmosphere that AI forecasting models are now hungry to see in much finer detail. Skyfora, a Helsinki startup, thinks the missing sensor network has been hiding in plain sight the whole time: the cell towers you walk past every day.

On June 9 the company said it had raised 6.5 million euros to scale a software layer that turns telecom infrastructure into a live atmospheric observation grid. No new hardware on the masts. No fleet of probes. Just clever signal analysis running on equipment that mobile operators already paid for and already maintain.

The round was led by a group of strategic and impact-focused backers including Eviny Ventures, Ugly Duckling Ventures, LUMO Labs, and the EIC Fund, with non-dilutive support from Business Finland. Small by megaround standards. Pointed, in a way that matters more than the headline figure.

The Sensor Network Nobody Had to Build

Here's the trick. Telecom networks are stuffed with GNSS receivers, the same satellite-positioning hardware that keeps base stations synchronized. Those signals travel down through the atmosphere before they arrive, and water vapor bends and delays them along the way. Measure the delay precisely enough and you can read the humidity in the column of air above the tower.

Skyfora processes those signal delays into real-time humidity readings, then streams them as weather data. Where a network lacks usable GNSS hardware, the company drops in its own StreamGNSS units to fill the gap. The result is a high-frequency, high-resolution view of the lower atmosphere across whatever footprint an operator already covers.

"We're turning existing mobile networks into the data layer for next-generation weather forecasting and climate intelligence," said Fredrik Borgström, Skyfora's CEO. The pitch to operators is almost cheeky in its simplicity. You own the towers. You maintain them anyway. Now they can throw off a second product without you lifting a wrench.

That framing is doing a lot of work. Telecom margins are thin and capex-heavy, and operators spend most of their energy defending revenue rather than inventing new lines of it. A software-only feature that monetizes assets already on the balance sheet is exactly the kind of thing a CFO will actually return a call about.

Why AI Models Are Starving for This

The timing isn't an accident. A new generation of AI weather models has arrived, and they're voracious. These systems learn patterns from enormous volumes of observational data, and their accuracy climbs with the density and frequency of what they're fed.

Traditional observation infrastructure can't keep pace. Most of the atmosphere stays under-observed, and the gear that does exist was never designed for the resolution modern models crave. So you get a strange mismatch: the algorithms have outrun the instruments feeding them.

Skyfora is selling the missing input. Dense, frequent, hyperlocal humidity data, harvested from infrastructure that already blankets populated regions. For a forecasting platform trying to call a thunderstorm three hours out instead of one, that extra granularity is the whole game.

The company is already working across several countries in Europe, the US, Africa and the Middle East, with the fresh capital aimed at deeper commercial deployments alongside telecom operators and forecasting partners. You can read the company's own framing in its funding announcement.

A Climate Bet Wearing a Telecom Suit

Strip away the GNSS jargon and you're left with a climate-resilience story. Extreme weather is getting more frequent and more expensive, and the industries most exposed to it, from energy to logistics to insurance, are desperate for earlier, sharper warnings.

Skyfora frames its data as critical national infrastructure, the kind of thing that helps a power grid prepare for a storm or a city brace for a flash flood. That's not just marketing gloss. It also explains the investor mix, which leans toward strategic and impact capital rather than pure financial returns. Eviny is an energy company. Business Finland is a public agency. These are backers with a stake in the resilience outcome, not only the cap table.

There's a quieter strategic angle too. An operator running Skyfora becomes, in effect, a steward of a national weather-observation network. That's a brand and ESG story telcos can tell regulators and customers alike. Sticky, defensible, and hard for a competitor to replicate without the same physical footprint.

Detail

Figure

Round size

6.5M euros

Headquarters

Helsinki, Finland

Lead investors

Eviny Ventures, Ugly Duckling Ventures, LUMO Labs, EIC Fund

Non-dilutive support

Business Finland

Core technology

GNSS signal-delay analysis on telecom hardware

Hardware required

None on existing towers; StreamGNSS where needed

Active markets

Europe, US, Africa, Middle East

Finland's Quiet Habit of Selling the Infrastructure Layer

There's a pattern in Finnish deep tech that's easy to miss if you only watch the consumer headlines. The country keeps producing companies that sell the unglamorous middle layer, the plumbing everyone needs and nobody wants to build twice.

Think about it. ICEYE built radar satellites and now sells Earth observation as a service. Nokia spent decades turning networks into platforms. Skyfora fits the lineage neatly: take a piece of installed infrastructure, find the latent data inside it, and sell that data to whoever needs it most. The capital efficiency is the whole point. You don't pay to manufacture and launch anything. You pay to read what's already passing through.

That model travels well, too. Cell towers exist on every continent, and humidity matters everywhere from Helsinki to Lagos to Riyadh. Skyfora's early footprint across Europe, the US, Africa and the Middle East suggests it understands that the real prize isn't Nordic weather. It's a global atmospheric data product priced as software.

What 6.5 Million Buys When Your Margins Are Software

A round this size wouldn't get a hardware company through a single prototype cycle. For a software-layer business, it's plenty. Skyfora's costs sit in engineers, signal scientists, and the commercial team that has to walk operators through deployment. No factories. No bill of materials that balloons with every unit shipped.

That's the leverage. Every new operator who switches the feature on expands the sensor network at near-zero marginal cost to Skyfora. The data gets denser, the AI models that buy it get hungrier, and the whole thing compounds without a matching spike in spend. If the company can string together a handful of marquee telecom names, the unit economics start to look genuinely rare for a climate-adjacent business.

The risk lives on the demand side, not the supply side. Forecasting partners and weather-exposed industries have to actually pay for higher-resolution data, and they have to keep paying once the novelty wears off. Prove the data sharpens a forecast in a way someone will write a check for, and Skyfora has a business. Fail to prove it, and you've got the most elegant science project in Helsinki.

The Catch Sitting Inside the Cap Table

None of this works without the operators. Skyfora's whole model depends on convincing telcos to flip on a feature, share signal access, and trust a startup to run analytics on their networks. That's a long enterprise sales cycle dressed up as a quick software toggle.

And the company isn't alone in chasing weather data from non-traditional sources. Crowdsourced sensor networks, connected cars, and other GNSS-based approaches are all circling the same gap. Skyfora's edge is the telecom angle and the no-new-hardware promise, but edges in deep tech have a way of narrowing once a market proves real.

The 6.5 million euros buys runway to convert early deployments into reference customers. If a couple of major operators go live and a forecasting partner can show the data measurably sharpened its model, the flywheel starts turning. If the sales cycles drag, this becomes a familiar deep-tech story about brilliant science waiting on slow procurement.

What makes this one worth watching is the elegance of the insight. The infrastructure was already there. The signals were already arriving. Someone just had to notice that the noise telecom engineers spent years filtering out was, to a meteorologist, the signal. For more on Nordic deep tech turning physical infrastructure into data plays, see our coverage of ICEYE's billion-euro round and Reel's energy expansion.

Finland keeps producing companies that treat the dull, expensive, already-installed world as raw material. A cell tower is just steel and antennas to most of us. To Skyfora it's a humidity sensor that happens to also carry your phone calls.

Now comes the hard part. Turning a clever idea into a network operators can't imagine switching off.

Keep Reading