Top-10 Artificial Intelligence Startups in the Netherlands
Most people know the Netherlands (or Amsterdam as the Americans call it) as a place where you can find good techno music, legal weed, and a liberal attitude towards sex. Others think about windmills, tulips, and bikes. It’s a lesser known fact that this small country (it actually has a smaller population than New Yawk State) is also a European hub of startups, thanks to its progressive tax incentives, visas for skilled workers, diverse English-speaking population, and great location. While VC funding is lagging behind large European cities like London and Paris, numerous startups and a handful of tech giants have established a presence in Amsterdam.
The country has just released a roadmap to a national AI strategy, lagging behind numerous others already on the path of execution. The roadmap concludes the Netherlands has lost its momentum in AI R&D, and the country needs a structured approach to AI development to attract much-needed funding. We scoured Crunchbase for the ten most-funded Dutch AI startups.
|Pyramid Analytics||Business Intelligence||Amsterdam||41.5|
|Effect.ai||AI as a Service||Amsterdam||13.6|
|3D Hubs||3D Printing||Amsterdam||11.5|
Founded in 2008, Amsterdam startup Pyramid Analytics has raised $41.5 million in funding to develop an enterprise Business Intelligence (BI) platform that employs machine learning. Pyramid’s unique selling point against its BI competitors is its flexibility. While other platforms tend to replace and mold existing infrastructure to fit their requirements, Pyramid adapts to existing systems like SAP, IBM, or Oracle, consolidates fragmented data sources, and uses existing infrastructure. The company’s Analytics OS product provides point-and-click interfaces that can prepare and deploy machine learning models, connect them, visualize results, and create presentations from the output. Deployment can be on the cloud, local, or a hybrid of both. Analytics OS is used in diverse industries including financial services, retail, healthcare, and manufacturing by companies like Siemens and Victorinox. The software’s pared-down Community Edition is free to download, and paid subscriptions start at $10 per user per month.
Update 03/27/20: Pyramid Analytics has raised $25 million in Series C funding to expand and to deepen strategic alliances. This brings the company’s total funding to $66.5 million to date.
We covered GeoPhy’s AI-based real estate valuation platform in our recent article on “Using AI for Commercial Real Estate Valuations.” Founded in 2014, the Delft startup has raised $33 million from big-name investors like Index Ventures early this year to create a commercial real estate valuation service using AI. Real estate valuations are tricky because, traditionally, these were provided by real estate agents based on comparable assets, but these same agents were incentivized to close a sale quickly and bag their commission. GeoPhy uses a variety of datasets, including satellite imagery, to analyze real estate prices on the fly using AI.
The valuation database is updated daily, valuations are instantaneous, and precise to a 5.85% deviation. Compare this to traditional appraisals which take weeks and differ from actual transaction prices by twice as much. GeoPhy’s algorithms also provide more granularity to the drivers of prices taking into account factors such as pollution, transportation options, age, students, and crime. The startup’s clients include names like Goldman Sachs, UBS, and APG, the Dutch pension fund.
Founded in 2008, Nieuwegein startup Flytxt has raised $22 million to develop an AI-driven marketing automation platform called NEON-dX. The product suite transforms customer data into real-time insights and analytics that can be used to personalize customer experience, deliver marketing campaigns across different channels, and automate marketing efforts across an organization. Flytxt’s algorithms promise to constantly optimize ongoing campaigns and highlight new opportunities, making inexperienced executives the stars of the trade. Like most startups, they’re looking to raise more money, and CEO Vinod Vasudevan expects the business to grow tenfold in the coming five years followed by an IPO. Flytxt already boasts numerous big-ticket clients including Vodafone, Samsung, and Volkswagen.
Founded in 2010, Amsterdam startup Mint Solutions has raised $15.1 million to develop computer vision technology for correctly administrating medication. Joining the 13 computer vision companies we covered in 2016, the startup has created MedEye, a scanning system that is able to verify that medication is dispensed correctly. According to the startup, one in five medicine administrations are wrong in one way or another (wrong patient, wrong time, wrong medication, wrong route, wrong dose) and half of these errors can be caught at the bedside.
MedEye integrates with hospital systems to highlight medications that are due, scanning tablets and oral solids to ensure all five criteria are right, and does this all using a webcam. It also takes care of documentation, helping to free up nurses to do more value-added tasks. The system can also inform patients and relatives of the medication administered and publishes analytics to doctors. MedEye has seen rapid adoption and has been implemented in 10% of Dutch hospitals in the past two and a half years.
Founded in 2015, Amsterdam startup Effect.ai has raised $13.6 million in an Initial Coin Offering (ICO) to develop a decentralized platform for AI development and the creation of various AI services. The team’s ambition is to “define the future relationship between humanity and AI” with their three-step project.
In phase one called Effect Force, the company has already established a marketplace for humans to train AI algorithms, much like Amazon’s Mechanical Turk. Freelancers can mess around with mundane data tasks and get paid in Effect.ai’s own cryptocurrency tokens called EFX. Phase two, the Effect Smart Market, will be a marketplace where AI and machine learning algorithms can be bought and sold as a service using EFX tokens. Phase three, Effect Power, will provide a decentralized computational framework that’ll run the most popular AI and deep learning models, also paid for by EFX tokens. While the first phase of development is ready, the other two are still in the pipeline with no due dates in sight. With the ICO closed, only time will tell if the team can deliver on their ambitious promises or fail like the majority of ICOs do. (If you’re dumb enough to fund a company without asking for equity in return, you get what’s coming to you.)
We last covered 3D Hubs’ distributed 3D printing business in our 2017 article on “3D Printing and Distributed Additive Manufacturing.” Founded in 2013, the Amsterdam startup has raised $11.5 million in funding to develop a distributed manufacturing platform, and since then has expanded beyond 3D printing into injection molding and CNC machining as well. Clients can upload designs of plastic or metal parts, specify requirements, and have the parts delivered to any address. The company provides instant quotes on their website using machine learning algorithms that utilize data from over 1.5 million previously sourced parts to provide highly accurate estimates of both time and cost.
It used to take 48 hours to provide estimates for parts as part of a manual process, but with AI these quotes are now fully-automated, instant, and far more accurate. A blog post by one of the company’s AI specialists talks about how critical machine learning has become to their business and that this will be the future of digital manufacturing. It’s interesting to see how the company continues to pivot – from 3D Printing community, to 3D printing bureau, and now to AI-powered manufacturing on demand.
Founded in 2010, Groningen startup Crowdynews has raised $7.1 million to develop a content curation platform using AI and Natural Language Processing (NLP) that recognizes user-generated social media content and pairs it with a publishers’ own content. The tool lets publishers gather and match real-life user stories with their own articles to boost their relevance and engage their audience. The platform uses and mixes different social media sources, filters unwanted topics, highlights trends, and does this in real-time. Crowdynews offers specific packages for sports publishers, adding live social content to match commentaries and news, and travel publishers, adding pictures, videos, and travel testimonies based on geolocation. (Nothing like taking a slew of typo-ridden tweets from a bunch of nobodies, slapping them into an article, then calling it “journalism.”)
Founded in 2014, Amsterdam startup Connecterra has raised $6.7 million to develop a dairy farmers’ assistant tool called Ida. The app gathers data using sensors placed on cows’ collars, relays the data to the cloud through field access points and a base station, then analyzes the input to come up with herd insights, behavioral analytics, health statistics, and fertility information. It also tracks changes implemented on the farm and measures their impact and their effect on the cows.
Connecterra is working with individual farmers as well as large industry clients like Danone, and has several thousand subscribers. For a subscription fee starting at $2.80 per cow, per month, Ida promises to increase the efficiency of farm operations, raise the well-being of livestock, and make farms more sustainable at the same time.
Founded in 2012, Enschede startup SciSports has raised $3.5 million to develop data intelligence tools for professional football organizations (that’s soccer for all you readers in ‘Murica), and industry valued at $28.5 billion in Europe alone. Their Insight platform is a database of 90,000 active players equipped with an algorithm measuring the quality and potential of every player in the world. The AI-based SciSkill Index provides apples-to-apples comparison of all the players and has outperformed bookmakers with an average Return On Investment (ROI) of 9.4%. It is the main selling point of the database, helping football professionals in talent scouting, recruitment, and data analytics.
The company’s other offering is BallJames, an autonomous camera system tracking the movement of players, the ball, and referees in real-time. Sounds similar to a Danish firm called Veo which we covered in our recent article on smart sports stadiums.
Founded in 2012, Nijverdal startup SafeSize has raised $3.1 million to develop a shoe recommendation and fitting engine. Shoes have significant size inconsistencies, and SafeSize uses scanners to measure and visualize shoe insides for fit. Another 3D scanner measures customers in-store, and recommends the right shoes for customers taking into account the activity they will use it for and the personalized 3D measurements.
The startup’s algorithms have been taught with measurements from 12 million people and 1.5 million shoes and SafeSize claims using the service increases sales conversions by 62%, allows a shoe price increase of 37% and results in 21% fewer product returns. The service is used by major manufacturers like Nike and multi-brand stores like Zalando.
The top ten list above is surprisingly diverse both in terms of location and application, something we don’t see very often when we do these AI country profiles. These companies have also received a fair amount of money, almost double the funding of Hong Kong’s top ten AI startups. With a new AI roadmap laid out, more and more Dutch AI startups should be able to attract even more international VC dollars.
Despite what the pundits say, FAANG stocks (Facebook, Apple, Amazon, Netflix, Google) don't give you real exposure to AI. Read about 7 stocks that give you true pure-play exposure to AI in our guide to investing in AI healthcare companies, freely available to Nanalyze subscribers.