Multidimensional Analysis of Global AI Unicorns: How Do They Grow and Expand?
From 2010 to July 2019, 36 AI unicorns emerged in six countries. While some AI companies focus on vertical industries like auto, healthcare, cybersecurity, surveillance and entertainment, more are now expanding horizontally.
From 2010 to July 2019, 36 private companies joined the Global Unicorn Club, according to CB Insights. Among them are AI companies using machine learning as a core differentiator, selling AI software or building AI chips, according to the official definition.
From a geographic perspective, six countries are home to these unicorns, with the United States and China dominating the chart. The two countries, with 16 and 12 unicorns respectively, are now universally regarded as the global AI leaders in terms of their talent pools, the number of startups, financing activities and monetization opportunities.
Despite the current position, U.S and U.K startups are the early birds in AI, with startups becoming unicorns as early as 2014 and 2015. At the early stage, investors place a higher premium on talent and algorithm in evaluating startups. Home to most renowned research institutes in AI, these two countries happen to be the driving force behind global AI fundamental research. According to a 2017 Mckinsey Global Institute report, U.S. and U.K. research results in this field are the most influential by citation impact index.
Meanwhile, Israel, one of the most innovative countries in IT, contributed three AI unicorns. The country of 8 million is a long-time tech innovation center. With huge R&D spending, a large number of start-ups and a high proportion of engineers to its entire population, the small desert country has a thriving venture capital investment scene, according to BBC.
The list of 36 AI unicorns also included unlikely candidates from countries not usually renowned for their tech prowess. UiPath became the only Romanian AI company to have made it into the top 36. It started its business in 2005 and later moved its headquarters to New York in 2017. UiPath specializes in robotic process automation (RPA), enabling robots to perform normally human tasks in sectors such as document management, contact center, healthcare, finance and accounting, human resources, and supply chain management.
Preferred Networks, Japan's only unicorn that made the list, provides IoT-centric deep learning systems in transportation, manufacturing, and healthcare sectors.
The earliest AI unicorns focused on improving business efficiency and efficacy as well as accumulating data advantages. InsideSales.com, which joined the unicorn club in 2014, is an AI platform that seeks to improve sales performance and provide better buyer experience through sales forecasting, engagement tracking, among other means. UK company BenevolentAI, which joined the ranks of global unicorns in 2015, specializes in medical data mining, generating hypotheses about new drug and testing candidates by processing scientific data and research thesis. Avant, based in the states, leverages financial data and machine learning in online consumer personal loans and became a unicorn in 2015.
According to the CB Insights survey, it took three of the 36 top global unicorns only one year to grow into unicorns.
Uptake, which was founded in 2014 and specializes AI + IoT, applies predictive analytics to interpret sensor data for clients in mining, energy, rail, aviation, retail and construction industries. As one of the early unicorn club members, the startup also has demonstrated its solid data capability and business productivity emphasis. Chinese company iCarbonx is also a "one-year" unicorn. It digitizes and computes personal gene and health data to meet personalized health goals. Founded in 2015, this Tencent-backed startup became a unicorn in 2016. Processor chip startup Cambricon, founded by scientists from the Chinese Academy of Sciences in 2016, became the first unicorn in its industry after its Series A round of financing in 2017.
Startups that became unicorns relatively recently tend to come from autonomous driving. Argo AI is dedicated to developing autonomous driving software in partnership with Ford and Volkswagen. Founded in the U.S. by Chinese entrepreneurs, TuSimple develops self-driving trucks in the logistics industry. Horizon Robotics focuses on AI processors and the integration of hardware and software. It has partnered with famous car makers including Audi, Bosch, Chang'an, BYD, SAIC Motor and Guangzhou Automobile Group.
Among the 36 AI unicorns in question, ByteDance seems to be a big whale swimming alongside small minnows, as it boasts a valuation of USD 75 billion after its pre-IPO funding round -- almost equal to all the others combined (USD 76.24 billion).
Beijing-based ByteDance is a content recommendation platform largely based on machine learning algorithms. Founded in 2012, its breakneck growth and global expansion has turned it into one of the most valued unicorns in the world. Currently, ByteDance has 700 million DAUs, mainly contributed by news portal Toutiao and short-video platform Douyin (TikTok). It reportedly hit sales revenue of USD 7.2 billion in 2018, with advertisement as a major source of revenue stream. However, these alluring figures belie the lingering doubts surrounding the key question of ByteDance's profitability, and we are expecting the possible IPO to unveil its financial picture.
Most of the AI companies in the CB Insights list are tech enablers in multiple vertical industries that range from auto and healthcare to cybersecurity, from surveillance to entertainment.
They mainly have strengths in one aspect of artificial intelligence - such as computer vision, voice recognition, robotic processing automation (RPA) and machine learning - that could be deployed to several vertical industries to improve productivity, increase accuracy or lower costs.
These companies normally started out in one vertical industry to leverage technical expertise. As they grow, startups are forming off-the-shelf availabilities and tapping into new monetization opportunities in other fields.
Indeed, AI knows no boundaries. In the future, more traditional companies employing this new technology could increasingly find themselves pushing the envelope in the application scenarios of AI and further blurring the boundary.