Nanalyze

Top-12 Artificial Intelligence Startups in Sweden

They say that when a blonde moves from Sweden to Norway, the average IQ of both countries goes up. Of course, the Swedes can get away with making such jokes because they’re the biggest country in Scandinavia – by far – with around 10 million people. That’s almost twice that of Norway. Or Denmark. Given her relative size in the region, it’s no surprise to see that Sweden has joined the global AI race as of this May with the release of its National Approach for Artificial Intelligence. The document will act as a reference for upcoming government policies and argues that the country needs to train more AI professionals, increase research efforts in the field, and develop a legal framework that will ensure the development of sustainable AI.

The report was released on the back of research from Vinnova, the government’s innovation agency, regarding Sweden’s development potential in the field. Since the government report, the state has started rolling out new policies regarding training funds, the establishment of a science park, and AI-related innovation projects driven by Vinnova. The end goal of policymakers is to increase the country’s competitiveness and to support its (already strong) welfare initiatives. In order to understand the sort of AI capabilities the country has presently, we found a list of the most funded AI startups in Sweden for you – courtesy of Berlin-based Asgard Capital, a VC specializing in AI investments globally:

Name Application City Funding (USD millions)
Seal Software Contract Discovery Gothenburg $43 million
Starcounter Software development Stockholm $27 million
Noomi Healthcare Stockholm $7.5 million
Strossle Publishing Stockholm $5.7 million
Watty IoT Stockholm $4.7 million
Univrses Machine vision Stockholm $2.7 million
Furhat Robotics Robotics Stockholm $2.6 million
Imagimob IoT Stockholm $1.6 million
Lytics Healthcare Malmo $1.1 million
Century Analytics Fintech Malmo $0.8 million
Loop54  E-commerce Stockholm $0.6 million
Refind Technologies Process Automation Gothenburg $0.12 million

The AI startup ecosystem looks very similar in terms of size and funding to its neighbor, Finland, the last country we covered in our AI country series. Let’s take a closer look at each of the top Swedish AI startups by funding.

The Top Artificial Intelligence Startups in Sweden

Click for company website

Founded in 2010, Seal Software has raised $43 million to create an AI a tool for automatic contract discovery and analysis. The company has been incorporated in the UK and has moved on to Walnut Creek in California since then, but is still rooted firmly in the Nordics with most of the development work located there. The company’s C-Suite offering uses machine learning and natural language processing to digitize, understand, and analyze contracts of clients automatically.

Seal Software contract tool

Source: Seal Software

The service does everything a capable assistant should, and much quicker. Duplicates are eliminated, contracts are searchable, and key information is mapped, extracted, and ready for analysis. Used by diverse industries like legal, life sciences, and finance, Seal Software has numerous big-ticket clients like Dell, PayPal, Salesforce, and Bosch.

Click for company websiteFounded in 2006, Stockholm startup Starcounter has raised $27 million to develop an in-memory database engine and application server for high-performance business applications. Powered by machine learning, the technology enables enterprise software developers to create apps that are able to share data and user interfaces with each other without the need for manual integration (a costly and time-consuming process). Automatic integrations allow this enterprise software ecosystem to operate as an app store for enterprise clients rather than each provider developing huge chunks of front-to-back software. Clients can mix and match best of breed apps without the need for complex and expensive integration projects. Besides its Stockholm headquarters, Starcounter has offices in Poland and Silicon Valley.

Click for company websiteWe’ve covered Noomi (rebranded recently from its original name Aifloo) a year ago in an article titled “Robotics and AI Assist in Caring for the Elderly.” Founded in 2015, the startup from Stockholm has raised $7.5 million to develop a healthcare system for the elderly that monitors behavioral patterns, analyzes the data, and allows caregivers to proactively address health risks and issues. The product comes in the form of a wristband equipped with a multitude of sensors which can monitor eating habits, sleep patterns, and trips or falls. AI algorithms analyze the input data over the cloud, and also alert caregivers in case of an emergency. Noomi partnered with leading Swedish healthcare provider, Aleris, in 2016 to develop and refine its product, and it is now used by the company’s elderly clients.

Click for company websiteFounded in 2013, Stockholm startup Strossle has raised $5.7 million in funding to develop an AI-based content recommendation engine for publishers. Facebook and Google currently generate more than 70% of traffic for publishers at the moment, and account for 99% of online advertising growth. Strossle has built a platform of more than 1,500 publishers that aims to take on these two behemoths. The team’s widget, powered by its predictive AI algorithms, recommends consumers relevant content at the bottom of an article based on previous user engagement, interest, and trending topics.

Topics consumed versus published in 2017 by Strossle

Topics read vs. topics produced by publishers, ratio, 2017 Credit: Strossle

According to the company, this results in 120% more time spent on a site and 40% more articles read compared to users coming from social media. Advertisers like Coca-Cola or Santander Bank also enjoy the predictive algorithms targeting very specific focus groups in the form of native content – those “sponsored articles” you see popping up occasionally. Strossle has been growing close to 100% every year since its launch and posted revenues of $11.4 million in 2017.

Click here for company websiteFounded in 2013, Stockholm startup Watty has raised $4.7 million in funding to develop a power usage monitoring platform for residential homes. The Watty hardware sensor connects to the fuse box detecting and identifying household appliances by their consumption patterns, something we’ve come across before. The integrated mobile app allows users to fine-tune electricity usage and save on the monthly bill, and also detects anomalies at home, for example when the fridge stops functioning.

The Watty power sensor and mobile app

Credit: Watty

Watty partnered with Chilean sustainable energy company Solarix a year ago to help with Chile’s drive towards sustainability and energy efficiency. The project involves equipping 1,000 homes in Chile with the Watty device. The first generation of Watty sensors have sold out and the company is accepting pre-orders for the next generation of devices that are under development.

Click for company websiteFounded in 2015, Stockholm startup Univrses has raised $2.7 million to develop precision motion tracking tools for the automotive, robotics, drone, and AR/VR industries. The team originally started working on AR/VR solutions including mobile positional tracking that has remained an unsolved issue until 2016 limiting the potential of VR on mobile platforms. With proper positional tracking for mobile, users are able to move in virtual space by moving in reality without a tether attached, or without mapping the real environment first. The company has since then expanded into the automotive, robotics, and drone industries and has developed algorithms for 2D and 3D object tracking, object recognition, and visual odometry – all of them essential in the fields above.

Click for company websiteFounded in 2014, Stockholm startup Furhat Robotics has raised $2.6 million to develop a life-sized robotic face capable of social interactions. Essentially a voice assistant with expressions, the Furhat robot has natural language understanding in 40 languages and an interchangeable face with the ability to emote according to the conversation. Here’s a video with an “it puts the lotion on its skin” feel to it, showing how Furhat works:

Furhat is offering the robot to businesses and organizations like airports, malls, hospitals, or offices with a standard face and set of emotions which can be expanded by the provider, or customized through a software development toolkit. At the moment, Furhat is far from the uncanny valley, but it does seem to provide a natural interaction experience to users. Already employed by global organizations like Intel, Honda, and Disney, it’s not inconceivable you’ll find Furhat running your next HR meeting. Maybe then, people will start taking HR more seriously.

click to company websiteFounded in 2013, Stockholm startup Imagimob has raised $1.6 million to develop AI algorithms combined with the Internet of Things sensors for mobile devices. The offering dubbed SensorBeat allows advanced motion detection for smaller IoT articles of virtually every kind. The software runs on low power processors and is autonomous, so it doesn’t need an internet connection or a central processing hub to work. It provides real-time insights and only sends data outside the equipped device if it’s necessary (for example, when caretakers of the elderly are alerted of a fall). This means a higher level of data security as well. Imagimob has already implemented its technology into smartwatches, steering wheels, industrial manufacturing lines, and power tools, among other devices.

Click for company websiteFounded in 2011, Malmö startup Lytics has raised $1.1 million in funding to develop big data analytics for healthcare. The company is offering different modules ranging from data capture through predictive models to simulations defining the right medications and setting the right dosage.

The Lytics medical big data project implementation

A typical Lytics project implementation, Credit: Lytics

Currently, these models are being used in conjunction with end-stage renal disease treatment, cardiology, and clinical trials. According to a study conducted by the company in U.S. health clinics, using Lytics systems translates to savings of 10,000 USD per patient per year for an institution. From the patients’ perspective, the system leads to lower hospitalization rates with a lower average number of days spent in the hospital. Lytics partners include major global names like Sanofi, Sony, and Boston Scientific.

click to company websiteFounded in 2015, Malmö startup Century Analytics has raised $800,000 in funding to develop AI algorithms for currency trading. The company is an asset management firm set up by an ex-trader and poker player, a former employee of Goldman Sachs, and a tech entrepreneur with AI and machine learning experience. Based on big data analytics, behavioral theories, and game theory, Century’s algorithms will power their first hedge fund to be launched by the end of this year. The company will offer its product to institutional investors and high net worth individuals.

Click for company websiteFounded in 2010, Stockholm startup Loop54 has raised $600,000 to develop a machine learning-assisted search engine specifically for e-commerce. The team has built a neural network that matches each product to a “neuron” that contains weighted features of the product. When enough feature weights are similar, a match occurs. Early searches rely on text matching, and as the platform learns over time, this is complemented by behavioral matches based on relevance. To put it simply, a search for potatoes will return fresh and processed potato products early on, and later the engine will complement these with turnips, parsnips, or celery as these carry the closest features to potatoes based on buyers’ behavior.

Loop54 Engine

Source: Loop54

Loop54 claims the engine results in a 30% increase in sales from search with an implementation timetable of 2-3 weeks. The search engine is used by Nordic clients of all sizes – even global chain Office Depot – with the average client spending $30,000 per annum on average resulting in annual recurring revenues of $3 million for the company.

Founded in 2008, Gothenburg startup Refind Technologies (previously known as Optisort) has raised $120,000 to develop a computer vision-based sorting solution for the recycling industry. (This reminds us of a certain Barcelona startup we met with this year.) Refind’s machines use cameras combined with machine learning to precisely grade and sort electronics, and determine whether they are good for reuse, refurbishment, or recycling.

Refind offers sorting and refund machines for used batteries as well, and, as a bit of a non sequitur, a species grading machine for fish. (Makes sense, since Sweden has a sizeable fishing industry.) The company is branching out and accepting projects to identify and sort anything that can be recognized by its looks. Refind offers its electronics sorting machines on a revenue-share rental basis, asking for a cut of the recycling revenues, and is already working with waste management and recycling service providers in Europe and the US.

Conclusion

This list contains a healthy mix of startups from different industries, but the only area that seems missing is education, considering that the Nordic education model is considered to be one of the most efficient and progressive in the world. Sweden may not threaten China’s leadership position anytime soon in the space of AI, but some good-natured competition among the Nordic countries will surely boost the sector’s output in the region. Hopefully, the new government initiatives will start having an effect, and more capital will flow into this interesting collection of Swedish AI startups.

Worried that AI might steal your job? You should be. Why not get a Masters in Machine Learning from London's Imperial College, one of the top-10 universities in the world? And you can study online!

Intelligence that’s not artificial

Let our MBAs stimulate you with insights once a week.