The Top-10 Artificial Intelligence Startups in Korea
Whenever someone talks about Korea, it’s safe to say they are referring to “South Korea”, but that’s not fair to the 25.5 million North Koreans who also consider themselves “Koreans” as well. (Also called the Democratic People’s Republic of Korea (DPRK), there are more people from North Korea on this planet than there are Australians.) When we talk about artificial intelligence in Korea, we therefore refer to the entirety of Korea, both North and South. We took a look at the DPRK’s technical prowess by having one of our MBAs go over there and poke around a bit. It wasn’t pretty. No hot water at times, electricity brownouts, antiquated weaponry, and run-down factories using outdated technologies:
While the people were lovely as can be, they don’t have a McDonalds, which violates our “nobody can have AI capabilities without access to a McDonalds” rule. To top it all off, we met a karaoke singer who didn’t know who Elvis was. That’s where we draw the line, so we moved on down to “the other Korea” to see if things were any better.
AI in South Korea
Home to huge tech conglomerates like Samsung, LG, and Hyundai, South Korea showed their commitment to growing AI earlier this year by announcing a $2 billion investment program to strengthen AI research in the country. The aim is to join the global top four nations in AI capabilities by 2022 with the establishment of at least six new schools focusing on AI and the training of more than 5,000 engineers. The government push was announced in response to China’s growing influence and the USA’s incumbent position globally. In order to see the top funded AI startups in Korea, we used Crunchbase. Yes, we recently complained about how bad CrunchBase sucks sometimes but if we miss a startup it’s much less likely they’ll complain because Koreans are just too darn polite. Here’s the list we came up with using CrunchBase data:
Let’s take a closer look at each of these Korean AI startups.
Four Disruptive Technologies in One
Founded in 2015, Seoul startup DAYLI Financial Group has raised $97 million to create a broad AI fintech platform that includes tools for predictive analysis and modeling, a robo asset management advisor, a trading algorithm, and blockchain technology. (They managed to use four disruptive technologies in one sentence if you’re counting.)
Besides its core businesses, DAYLI also has a slew of subsidiaries offering complimentary tools loosely related to its core offering. The company acts as a one-stop-shop for everything related to fintech and is present in Asia Pacific, Europe, and North America. The applications are built as separate modules and are available to financial institutions separately so clients can build their own systems as if using Lego blocks. Use cases show significant business value-add, for example a 10% increase in loan approval rates, or credit card user growth of 20% post-implementation.
Improving the Precision of X-rays
Covered in our article “12 Startups Diagnosing Medical Images With AI”, Seoul startup Lunit was founded in 2013 and has raised $20.5 million to develop a medical imaging algorithm detecting chest and lung abnormalities. The application can screen for lung nodules including lung cancer, breast cancer, and other major lung diseases like tuberculosis. Since our 2017 article the company, has raised an additional $15 million from investors including Softbank and has launched Lunit Insight, a wholly cloud-based real-time x-ray analysis solution that has been used to analyze 1 million images across 80 countries since its launch.
Lunit’s algorithms aim to increase the precision of x-ray diagnostics that are 20% below that of CTs. Most early examinations use the cheaper and less precise x-ray method that leaves room for false negatives. The company is planning expansion to Europe and the US with regulatory approvals expected in 2019 and 2020 respectively.
The Netflix of Korea
Founded in 2011, Seoul startup Watcha started as a movie rating and recommendation app, and has raised $19.6 million in funding so far. Called by some the “Netflix of Korea”, the company has since grown into a VOD (video-on-demand) service, and (unsurprisingly) a competitor to Netflix. The recommendation app, based on users giving ratings to at least 15 movies, is now a major selling point for the streaming service with a database of 350 million ratings for hundreds of movies. The VOD platform subscription, called Watcha Play, costs about half of what Netflix charges at $4 per month, and according to reviews, has more diversity in movies (100% Asians we hear.) Netflix is stronger in recent drama series though, and critics highlight the recommendation functionality as Watcha’s main selling point.
Getting Rid of Your Engrish
Founded in 2014, Seoul startup Riiid has raised $13.3 million in funding to develop an AI platform that helps students study for the internationally accepted English language exam TOEIC. The app called “Santa TOEIC” is an adaptive learning solution that lets students practice the topics most relevant to them by analyzing their learning behavior and offering practice questions accordingly. The AI algorithm also calculates the easiest route to scoring more points on the exam, optimizing the time invested in studying. The company is planning to implement this custom learning curriculum approach to other exam types like the TOEFL English language exam or the SATs.
A Little Bit of Everything
Founded in 2015, Seoul startup Skelter Labs has raised $10 million to develop AI technologies both for industrial and retail use cases. The core technologies Skelter Labs is working on include natural language processing (NLP), machine vision (for manufacturing quality control), speech recognition, and context recognition from multiple sources (for example GPS, sound, or calendar events).
These technologies can be applied to smart devices, automobiles, wearables, or even industry verticals, and founder Ted Cho confirmed numerous projects are in the pipeline including a “deep-learning project for building AI technologies that match high-level cognition”. The company is secretive about its clients, but word on the street is they are working with industrial corporations while developing their own retail hardware offering.
Getting More Drugs Faster
Founded in 2015, Seoul startup Standigm has raised $3.7 million in funding to develop an AI platform for drug discovery and development. The development of new drugs is a long and complicated process starting from the discovery stage going through preclinical and clinical research ending in regulatory approval and safety monitoring. The process can last more than a decade with a final cost of $2.9 billion even if successful. The success rate of Phase I trials (already some way down the line) to drug approval is 13.8% on average. Standigm uses machine learning on existing biomedical databases to eliminate some of the uncertainty from the drug discovery process and find the most promising candidates. Besides drug discovery services, the company is also running projects on exploring the relationships between drugs, proteins, and diseases, predicting synergies between multiple drugs, and predicting new indications for existing drugs.
Making Factories Great Again
Founded in 2011, Seongnam startup ulalaLAB has raised $2.6 million to develop smart factory solutions combining Internet of Things (IoT) sensors with AI analytics making sense of IoT sensor data. The company’s WimFactory platform targets SMEs and equips manufacturers with sensors collecting data on the temperature, moisture, and pressure of facilities, and allows them to monitor and analyze this data in real time.
Users are able to increase productivity, make quality control more efficient, and implement preventive maintenance measures. According to ulalaLAB, the smart factory market will grow to $339 billion by the end of 2028 and the company is on the right track to capitalize on this growth with $16 million worth of export deals signed recently to expand the business to Taiwan, Thailand, China, and Indonesia. High profile brands like Nike and Adidas are among the company’s clients in the region. ulalaLAB is working on a smart farm solution as well.
3D Replay Technology for Video Games
Founded in 2014, Seoul startup Minkonet has raised $2.5 million to develop a proprietary 3D replay technology for video games. Their algorithms are aimed at the $900 million e-sports market as well as streaming and social media sharing services, and provide 360-degree high-definition replay, enhanced social media interaction, and new content capabilities.
The technology first appeared in PlayerUnknown’s Battlegrounds and has received praise from gamers globally. The 3D replay allows players to cut and edit videos of their gameplay, and post it to social media sites. It also prevents cheating (or hacking), a problem that developers are spending millions of dollars to mitigate. In addition, it will support the placement of native advertisements in games, a high growth segment of the advertising industry. Minkonet has set up its North American headquarters in Los Angeles in 2017 and is planning to expand in the US.
There’s Money in Corpus
Founded in 2012, Seoul startup Flitto has raised $2.2 million to develop a crowdsourced translation platform. We covered the company in our article “7 Machine Language Translator Startups” earlier this year, and Flitto’s relationship to machine learning lies in their collected language data called “corpus” that it sells to large corporates like Baidu, Microsoft, and Tencent. Corpus is used to train machine learning algorithms that leads to precise automatic translations by AI. The lion’s share of Flitto’s revenues comes from selling corpus, and the division is growing at a fast pace. Sales volume projected for 2018 is 30 million data points versus 6.9 million in 2017. Most popular languages requested are English, Chinese, and Japanese, and the top categories are colloquial, travel & shopping, and IT & software. With the machine translation market expected to triple in the next six years, Flitto is in a great position to sustain leadership in the corpus space.
It’s Not a Chatbot, It’s “Conversational AI”
Founded in 2016, Seoul startup Moneybrain has raised $1.8 million in funding to develop a chatbot service of their own. Joining the growing number of chatbot apps, platforms, and frameworks out there, Moneybrain promises natural language processing, continuous machine learning based on previous conversations, and flexible APIs (programming interfaces) that connect it to chat platforms or internal systems for task execution. The chatbot is gauging speakers’ emotions, looking at context, and creating answers from scratch to arrive at the optimal flow for a conversation. The team is working on a speech synthesis technology to allow the bot to answer in natural speech as well. Moneybrain currently works with regional clients including Korea’s largest credit card company, Shinhan Card.
While researching the most funded AI companies above, we’ve excluded two that had taken in meaningful amounts of funding and were classified as “AI startups”:
- Urbanbase – received $4.3 million to create AR/VR solutions for interior planning and design.
- Nearthlab – received $2.9 million to develop an automated industrial inspection service for drones.
While both companies employ AI as part of their services, their core offerings would be classified under other categories we cover (AR/VR and Robotics respectively) so we exclude them from this list. (Maybe we’ll cover them later if they bribe us with a few cases of soju.) It seems like it’s all happening in Seoul these days, and with the government funding and educational push towards AI R&D, we hope to see more exciting Korean AI startups pop up, maybe in other major cities as well – like Pyongyang.
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.