With a population of around 24 million people, Taiwan is around the same size as Belgium with twice as many people. It’s a densely populated country with delicious night markets, garbage trucks that play Beethoven, and some of the smartest people on the planet. The dynamic capitalist economy is largely driven by industrial manufacturing and electronics exports. While the country has strengths in AI, advanced manufacturing, and robotics, its startup scene remains small because of the lack of entrepreneurial initiative, the lack of talent, and a tough regulatory environment. Taiwan’s current president, Tsai Ing-wen, has pledged to help startups by introducing tax credits, attracting foreign talent, and setting aside land for startups. Taiwan also released a four-year AI action plan in January 2018 with an annual budget of around $320 million that includes five key initiatives:

  • Train local AI talent and attract international AI professionals
  • Establish an R&D institute similar to DARPA in the US
  • Create an AI Innovation Hub that will foster 100 AI startups
  • Open up data flow and establish flexible regulations
  • Integrate AI technologies into the government’s larger industrial innovation initiative

In order to better understand Taiwan’s present-day capabilities, we scoured Crunchbase to find the Taiwanese AI startups with the most disclosed funding to date.

Name Application City Funding
(USD millions)
Appier Online Marketing Taipei         81.5
Viscovery Computer Vision Taipei         15.0
iKala Online Marketing Taipei         10.0
AIgatha Marketplace for AI Algorithms Hsinchu           2.9
GliaCloud Automated Video Generation Taipei           2.5
Starwing Indoor Positioning System Taipei           2.0
aetherAI Healthcare Taipei           1.1
WritePath Financial Translation Taipei           0.5
hiHedge FinTech Taipei         0.02

 

Click for company websiteWe first came across Appier in our article on “The Top 10 Artificial Intelligence Startups in China” a few years ago, and you may be tempted to ask why this company has been classified under “two countries”. This a great opportunity to let you know that we don’t use the “C word” when talking about Taiwan, and you shouldn’t either. What matters is that Appier has raised $81.5 million in funding to develop their AI platform that initially addressed cross-screen-marketing. Simply put, using that solution means that users are 40% more likely to click your ads. They’re also moving into other areas.

Appier recently released two new modules besides its cross-screen advertising tool. One of them is a big data product that predicts customer behavior, the other an audience engagement tool which uses AI algorithms to analyze virtual behavior. Appier has many global corporates as reference clients, and has helped cosmetics company AXE, among others, to expand its mobile audience by 50% through engaging existing mobile users on multiple devices and identifying new customer segments that behaved similarly to existing ones. Another cosmetics client, Estée Lauder managed to roughly halve their cost-per-click and cost-per-lead using the platform, and use cross-screen advertising on different devices to convert 11 times the leads compared to advertising on PC only. The company projects global mobile ad spend will reach $250 billion next year.

Click for company websiteFounded in 2013, Viscovery has raised $15 million to develop computer vision solutions for a number of industries. The startup’s algorithms can recognize objects, faces, and scenes, on photos as well as videos. Viscovery’s off-the-shelf products include self-checkout applications for bakeries (yes, the app recognizes different kinds of breads) and other retailers, retail analytics solutions that measure footfall and “shopping journeys”, video labeling and scene search for movie makers, and of course bog-standard facial recognition tools that they probably won’t be able to use in Silicon Valley soon.

Viscovery's retail analytics provide insights on the shopping journey including footfall and customer details,

Viscovery’s retail analytics provide insights on the shopping journey – Credit: Viscovery

The flagship self-checkout offering can be easily integrated into existing point-of-sale systems and promises to decrease checkout transactions to one second from an average of one minute, decreasing queues and increasing revenues. The company offers custom consulting for image recognition projects in various industries where their team will work to define the ideal training model and develop a tailored set of algorithms for clients. Viscovery has won numerous startup awards since its founding, including recognition from Google and the Taiwan Ministry of Science and Technology.

Click for company websiteFounded in 2011, iKala has raised $10 million to develop AI-driven online marketing products. The startup offers five distinct solutions. Its Shoplus suite supports product sales through Facebook with order management, payment, and logistics consolidated and managed by a chatbot. KOL Radar is a directory of influencers – also called Key Opinion Leaders or KOLs – in major markets of the Far East that matches the best influencers to brands using AI algorithms and social media analytics. StraaS is a video marketing platform that helps video marketers share and live stream videos and engage with audiences using these videos. CloudAD uses big data analytics to automate the advertisement processes of e-commerce websites on Google Ads. The last one, GCP.expert supports businesses running their online services on Google Cloud. iKala’s services are used by corporations inside and outside the Asia Pacific region in multiple industries including gaming, media, e-commerce, and finance.

Click for company websiteFounded in 2018, AIgatha has raised $2.9 million to develop a distributed computational network based on blockchain technology, that will at later stages become a marketplace for AI algorithms. The startup’s aim seems similar to Effect.AI in the Netherlands – to establish a central exchange to buy and sell AI services. The first stage of the project is already in progress – the startup is building a blockchain where cryptocurrency miners offer their computing capacity for AIgatha tokens, and users pay for this capacity using the same tokens. The second stage will integrate miners’ computing power into a distributed framework by Q1 2020, and the third stage will involve the creation of a marketplace for AI algorithms that run on the distributed network in Q2 2021. In the final stage of development, the startup intends to overlay this AI marketplace with a user-friendly graphical interface that will allow anyone to implement AI solutions easily into their business. The startup has released their key stage two developer library for distributed computing as open source on Github.

Click for company websiteWe first came across GliaCloud, an application using AI to convert text into videos automatically, a year ago. Founded in 2015, GliaCloud has raised additional funding from investors since our last article bringing their total funding to $2.5 million. GliaCloud’s AI algorithms analyze text files from URLs, create scripts from identified topics and keywords, and generate high-quality video clips with embedded highlights based on these scripts.

GliaCloud turns texts into video scripts, then videos automatically to offer engaging content

Credit: GliaCloud

The company claims their video generator can increase production volume tenfold, freeing up half the time of content creators who can then focus on more value-added tasks. Last year, GliaCloud partnered with Inmagine Group, the owner of one of the world’s largest curated content libraries, to provide high-quality stock photos and footage to clients. The service is used by major publishers in the Asia Pacific region.

Click for company websiteFounded in 2014, Starwing has raised $2 million in funding to develop a highly accurate indoor positioning system that tracks movement using positioning tags or smartphones, and analyzes indoor movements and behavior with AI algorithms. The company’s tracking devices are used in warehousing, healthcare, prisons, museums, shopping malls, and airports to track the movement of people and objects with centimeter-level accuracy.

Starwing's indoor positioning system can check and limit location of prison inmates and analyze their social behavior

Starwing applications in a prison – Credit: Starwing Technology

Starwing’s application can be used to help customers navigate in a mall, analyze the social behavior of prison inmates, or monitor hospital patient flow. The company offers its positioning solution in a package bundle that includes hardware, software, and custom installation, all of which are priced on a case-by-case basis.

Click for company websiteFounded in 2015, aetherAI has raised $1.1 million to develop medical image processing tools focused on digital pathology, the analysis of digital microscopic tissue slides. The startup’s web platform includes an image viewer, annotation tools, and a deep learning framework for image analytics. (Sounds like some of the work we saw on display at a Stanford medical imaging conference earlier this year). aetherAI has built a cancer screening model and prognosis prediction models using the platform, and is awaiting the launch of clinical trials. The company works with local and international hospitals including the Cedars-Sinai Medical Center of Los Angeles, the University of Pittsburgh Medical Center, and the National Taiwan University Hospital.

Click for company websiteFounded in 2009, WritePath has raised $525,000 to develop an AI translation service for the financial industry. The startup uses a mix of human translators and AI algorithms to translate official documents like annual reports, IPO prospectuses, and pitch books in banking, fund management, and insurance into 35 languages. (If they can actually translate Chinese financial reports with no Engrish, sign us up.)

The startup’s algorithms handle all redundant translation tasks like text repetitions, glossary terms, and previously translated segments while also managing and unifying terminology throughout a document. WritePath also provides an Application Programming Interface (API) that integrates with client systems and can fully automate the translation of web and mobile content. On-demand translations start at 0.32 cents per word with discounts offered for package deals.

Click for company websiteFounded in 2012, hiHedge has raised $18,000 to develop an AI trading platform for financial markets, something we’ve seen before in various manifestations. In the case of hiHedge, their algorithms ingest price and trading volume data – what they call market data – from exchanges around the world. Add to that data some news, macro trends, and company reports, and the algorithms can begin to find patterns in this big data and develop predictive models on their own. These predictive models then become trading strategies that match a client’s trading profile and evolve as the markets change. hiHedge’s models continuously optimize costs and risks, and notify users of investment signals in real-time. The company received its seed funding from Startupbootcamp, a global network of industry-focused accelerators focused on early stage investments.

Conclusion

There are most likely some Taiwanese AI startups out there that flew under our radar because they weren’t listed in Crunchbase with total disclosed funding to date. That’s usually because the founders are too busy executing – or too busy eating oyster pancakes and getting hopped up on baijiu. In either case, that’s totally okay. Read our article on how to market your startup using the least amount of time and money as possible, then drop us an email. We’d be happy to have a chat and see what you’re getting up to.

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!