Ignitia Bets on AI to Disrupt the $3 Billion Hazard Warning Market

Ignitia is a deep-tech company committed to democratizing weather and climate intelligence around the world. It provides AI-powered weather insights for adaptation and resilience designed to keep you ahead of the storm.

Using advanced satellite data and machine learning, the company delivers highly accurate, up-to-the-minute updates on thunderstorms, severe rain, dry spells, and other critical weather events — without the need for physical sensors.

This is the second in a three-part sponsored series on Ignitia. You can listen to our special podcast episode with Ignitia CEO Andrew Lala HERE.

Interested in being a Climate Proof sponsor? Get in touch HERE

Being able to predict the weather can be the difference between life and death. Throughout history, it’s also determined who gets fed and who goes hungry; who grows rich and who descends into poverty.

Today, climate change is exacerbating the gulf in outcomes between those with access to accurate, reliable, and prompt weather information — and those without. With weather extremes becoming, well, more extreme, the consequences of being unprepared are becoming ever more existential. 

Hence why experts project booming demand for climate and weather intelligence services. Earlier this year, consultancy BCG said it expected the US$3bn weather hazard warning market to grow around 10% per year for the next five years, thanks to a global shift to “industry-specific solutions, hyper-local forecasts, and predictive insights.”

It’s also why the UN is investing so much political and financial capital into Early Warnings for All — its plan for bringing timely hazard alerts to everyone on Earth by the end of 2027. The ‘returns’ in lives saved could be extraordinary, since the UN says countries with good early warning systems have a disaster-related mortality ratio nearly six times lower than those with patchy or non-existent alternatives.

These trends are fueling Ignitia — an AI-enhanced software and data provider delivering high-precision weather forecasts and climate intelligence to users around the world. From farmers in Africa looking to make informed decisions on planting and harvesting, to port operators searching for insights on how storms could snarl vessel traffic, Ignitia’s real-time weather alerts and advisory services are giving users an edge in a warmer and wilder world.

“We are democratizing weather and climate intelligence,” says Andrew Lala, Ignitia’s CEO. “We focus particularly on ‘last mile’ distribution systems for our customers, where we see a huge gap in the uptake of these types of solutions that can help build resilience.”

While the company got its start selling weather alert subscriptions to farmers in Ghana via text and WhatsApp, it’s now expanding to supplying data and forecasts to the lucrative financial institutions and insurance space, as well as to large infrastructure asset owners and local governments. 

International organizations are paying attention. Indeed, earlier this year Ignitia was tapped by the UN’s AI for Early Warnings for All Innovation Challenge, where promising climate intelligence companies work out how to partner with public authorities to provide effective, inclusive early warnings to the most vulnerable populations.

Ignitia made its first breakthrough achieving what most businesses aspire to do but few accomplish: finding a solution to a common problem.

When the company was founded 15 years ago, its brains trust was operating in Ghana, where it maintains an office to this day. Here, the tech team — led by Co-Founder and Chief Scientist Andreas Vallgren — was grappling with a stubborn forecasting challenge that was causing local farmers all sorts of problems: rainfall. 

“Global forecast models have very little skill in terms of rainfall prediction in West Africa. It’s one of the hardest places on Earth, actually, to predict the weather,” he explains.

At the time, one big issue with these global models was that their resolution was too coarse to resolve the type of localized rainfall common to this region — and others beyond it. Another challenge is West Africa’s proximity to the equator, where the mathematical equations useful for modeling weather patterns elsewhere on the planet break down.

But more than anything else, forecasting rainfall is complicated by the atmosphere’s inherently chaotic nature, and the range of interacting spatial and temporal scales involved. Myriad factors determine whether, when, and where rain will fall — from large-scale systems like the atmospheric jet streams, to minute shifts in droplet formation within specific clouds. Modeling them all is tricky.

In West Africa, Vallgren and team also struggled with the lack of ‘ground truthing’. Large swaths of the continent and other tropical regions do not have ground observation infrastructure — like radar and weather stations — in place, making it hard to validate the characteristics of remotely forecasted rainfall events. This also complicated Ignitia’s efforts to refine its models and produce valuable ‘nowcasting’ predictions of current and near-future events — the kind farmers hanker after.

Vallgren and team worked to overcome these challenges with a numerical approach, working step by tortuous step to improve the modeling of the physical processes that underpin the formation of rainstorms. They also worked to improve the resolution of global forecast models so that rainfall estimates could be rendered at a more local scale.

The outputs of these refined models were pushed to farmers in Ghana and elsewhere in West Africa over local telecom networks in the format they are most familiar with: phone and app messages. “ We have lots of tools for farmers, like rainy season onset, real-time alerts — when to plant, when to apply inputs, when to harvest,” explains Lala. “They don’t need convincing that better data improves yields or helps them anticipate pest shocks and or weather shocks — it’s a no-brainer for them.”

This is confirmed by the sheer number of those signed up to the company’s early-warnings service in West Africa: a total of 3.1 million to date, with around 500,000 monthly users.

About seven years ago, everything changed. With machine learning, a branch of AI, taking off across business and academia, Ignitia pivoted to a new, cutting-edge approach to predicting the weather.

Vallgren’s team started by asking tropical forecasters to share exactly what they were looking for when making short-term forecasts and the sources they relied on. These included satellite images and other indicators from ground- and space-based sensors. They then fed these datasets into a proprietary machine learning model, and found their forecasting accuracy improved dramatically. “2018 is when we really started launching the AI part of the company, and we have seen enormous success in solving real problems,” says Lala. 

Indeed, in the Global South an AI-first approach to advanced weather forecasting is not only a faster, cheaper solution — it’s the only one that makes sense for that part of the world, he argues. “As you look at the Global South, there is not the same density of observation data that we have in North America and Europe. We have tons of radar, tons of weather stations, tons of satellites. The Global South does not. And so the way we built the weather machine in the US and Europe is not the same way that the Global South is going to be building their weather machines or the way that we are going to build the forecast system of tomorrow,” he explains.

To be clear, Ignitia’s AI suite is a world apart from the Large Language Models (LLMs) that have swept the globe since GPT 3.5 launched in 2022. The company believes the inherent randomness and statistical illiteracy of popular chatbots is simply not a good match for weather forecasting — where accuracy is of the utmost importance.

“One of the big things we debated early on was how do we avoid a catastrophic miss and catastrophic false alarm,” explains Lala. “ We never will have a forecast that could be hallucinated and then it’s distributed.”

He can make such a promise because Ignitia’s models remain anchored in observation data sets, and when it comes to filling in the gaps with AI the company carefully tests the methodology in regions with similar climate patterns — and rich in ground observations — to ensure realistic behavior. In addition, it applies threshold conditions based on local climate norms. Alerts are only triggered when these conditions are met, helping to avoid awkward or unnecessary warnings: like sending a dry spell alert for the middle of the Sahara.

Another help has been the company’s embrace of a spin-off innovation from its AI work — a real-time ‘virtual radar’, which produces AI-informed rainfall estimates for the very near future. “By combining data from our real-time virtual radar with satellite based rainfall estimates, we’ve seen huge improvements in all sorts of error metrics. This helps produce both better forecasts but also allows for looking back at past events and building more accurate climate risk products,” says Vallgren.

The results have indeed been impressive. A third-party validation from 2019 shows Ignitia’s models beating the Global Forecast System, housed by the National Centers for Environmental Prediction in the US, on probabilistic rainfall forecasts in 93% of global regions between 50°S and 50°N. In the remaining 7% — primarily arid zones like the Sahara — the results are tied. Over 13 months of daily comparisons, Ignitia provided more accurate forecasts every single day. Similar outperformance has been measured relative to forecasts produced by the European Centre for Medium-Range Weather Forecasts (ECMWF), the global gold standard.

*Accuracy comparison of ECMWF probabilistic rainfall forecast vs the Ignitia probabilistic forecast at one day's lead time, based on 393 days of daily rainfall probability forecasts. Source: Ignitia

Still, this approach — tying forecasts to observations — has its challenges. In a pre-global warming age, what happened weather-wise in a certain part of the world at a certain time of year could be highly predictive of what might happen the next year. With temperatures rising and the climate changing, this is no longer necessarily true. Lala acknowledges the difficulty of keeping the right balance: allowing AI models enough freedom to account for growing uncertainty, while also keeping errors in check. 

While the company’s tech was initially created to serve West African farmers, it wasn’t long before Ignitia recognized that many customers, in many different kinds of markets, had an appetite for highly-accurate, super-fast weather alerts.

“Until recently, it was not very well known that there was this big problem with rainfall forecasts — and not only in tropical areas of the planet,” explains Vallgren. “It’s a challenge in the summertime in the mid-latitudes too, including in parts of the US and Europe. You can see a big drop in forecast accuracy in many areas at this time of year.”

Hence the company’s recent pivot to servicing customers in the Global North — particularly owners of financial and real assets who have to know if, when, and how badly they could be disrupted by extreme weather events. Today, over 12 multinational companies are piloting Ignitia’s tools in at least one geographic location or for one group of assets.

The financial services sector is of particular interest. This is ripe territory, crowded with the kinds of organizations, including insurers on the frontlines of extreme weather risk, that can and will pay top dollar for the best in forecasting intelligence. But because of this, it has already attracted a swarm of climate risks analytics start-ups, all hoping to be the one solution to rule them all. 

Lala says this kind of thinking is misguided, though, and is betting on a different environment unfolding in which companies like Ignita can thrive. “It’s not going to be the case that there is going to be one super-intelligent meteorologist that everyone’s going to rely on. That’s not how this is going to go. There’s going to be many different AI models to solve many different problems for different industries,” he says.

This outlook explains why the company is picking and choosing its spots carefully. Instead of rolling out a one-size-fits-all data API, Lala and his team are working to understand each sector and what Ignitia could bring to it. After all, it can’t simply jerry-rig what it’s built for African farmers for global reinsurers, or major US port operators. 

Translating its original capabilities for a new user set has — and continues to be — a highly involved process. “Traditionally, we’ve been selling access to our API, but in many cases we found that even if companies have access, they don’t really know what to do with the data, or it’s not informing any other systems on their side,” says Vallgren. “They want either a PDF bulletin that explains what to look for, or for us to provide a CSV file with the data so that they can play with it in Excel.”

While it’s more hands-on work expanding the customer base this way, Lala says it’s the only route to mainstreaming weather intelligence as a product class. The advent of LLM-powered prototyping tools and low-code software building packages has been a big help here, he adds, as they’ve enabled bespoke data solutions to be spun up faster than ever before.

“I think the driving reason why many businesses have not already adopted weather and climate intelligence products is because collectively we — solutions providers — have not been able to bring the right solution in the right format to them. And I think now is the moment where you’re seeing this kind of disruption because anybody who has access to better data than any other alternative is able to produce a superior product for customers. I think it’s a massive boost for climate adaptation and resilience solutions on the market,” Lala says.

While the company has big plans for financial institutions and other high-paying clients, it continues to explore how its tech can help the most climate-vulnerable communities — where early warning systems can quite literally save lives.

“We’ve always had in the company’s DNA this idea that we want to democratize solutions to 200 million individuals. They could be using our data or having apps powered by Ignitia, but it’s not enough for us just to knock it out of the park with sales. We want many people to benefit from this type of technology,” says Lala.

This mission underpins his enthusiasm for the UN’s Early Warnings for All (EW4A) initiative, and Ignitia’s selection by the initiative as a top four finalist in its Innovation Challenge. This was awarded for the company’s virtual nowcast radar, which is capable of delivering hyper-local storm alerts with up to three-hour warnings.  “It’s actually faster than a traditional radar, because to perform a complete radar scan, it takes around five to 10 minutes, and then there’s post-processing to consider. In our case, a nowcast can be delivered in two minutes, because the AI model itself runs very fast,” says Vallgren. 

The virtual radar is also cheaper to run than more traditional services — an important consideration for the EW4A initiative, which is working out how to deploy nowcasting solutions in some of the poorest regions on earth. Ignitia has so far reached over 30,000 users across Latin America, Southeast Asia, and Africa with its real-time nowcast alerts.

Lala points out that, to win UN recognition, the company didn’t have to launch fancy new weather satellites or invest in new supercomputers. It simply had to leverage its existing models and text alert distribution system. It’s a lesson he’s taken to heart: “ There are certain problems that you don’t need hardware to solve, and there are certain problems that partnerships are much more efficient and effective at solving because they’ve already made those infrastructure investments,” he says.

This more nimble, software-first approach is the one he and Vallgren trust will bring the company success as it expands its horizons. “We have only just scratched the surface with some of these models, and we’re really eager to see how this renewed interest in extreme weather risks by many of these businesses will help us push forward new solutions,” says Lala.

Thanks for reading!

Louie Woodall
Editor