Investments in cognitive technologies, Q1 2016: a broader view on the ecosystem
The US, the UK, and Canada
Cognitive tech is on hype. Nevertheless, it is still underestimated. Let’s look at the wider ecosystem of cognitive tech rather than only at the artificial intelligence (AI) segment. Some include in the AI segment ‘…startups primarily focused on developing AI, across areas including image processing, natural language processing, machine learning, deep learning, and predictive APIs, among other core applications’. However, cognitive tech ecosystem covers machine learning, natural language processing, optimization, planning and scheduling, robotics, rules-based systems, speech recognition, and compute vision.
Summary — what to watch
- It appears that the UK is driven by exceptional events like large M&As, large funding rounds, while the US works as a conveyer;
- One may expect the UK will keep up with the US in terms of early stage cognitive tech funding volumes;
- Looks like VCs allocate small budgets to those companies in cognitive tech which raise their first rounds (<$100M in Q1 2016);
- To look at vertical players, especially in Finance/Insurance and Healthcare where a lot of non-routine processes are challenged by cognitive tech;
- To look at leaders among horizontal players, who may act as acquirers;
- Expect unexpected acquirers, Q1 2016 shows that a buyer may pop up from almost anywhere.
- Widening the scope gives more data. Instead looking at 27 deals and $81M invested in AI during Q1 2016, at least 123 cognitive tech companies that raised more than $638M in Q1 2016 and nine exits during the same period (Chart 1) can be analyzed ;
- US companies consumed almost 88% of $638M invested in cognitive tech across three countries, namely the US, the UK and Canada. ~10% came to UK’s firms and less than 2% to residents of Canada;
- The UK is significantly far behind the US in terms of first round funding volumes than in later stage funding rounds (Chart 2). The narrower gap in later stage funding rounds is due to the large round of Blippar. As for first funding rounds, it seems that there is room for more early stage investors in the UK, the ones like Entrepreneur First, an active cognitive tech investor during Q1;
- Companies raising subsequent rounds drive volumes, while companies that raised their first and only rounds during Q1 2016 got 15% of invested capital. Moreover, those who raise their subsequent rounds in Q1 were lucky enough to attract ~40% of their total finding during that period. Therefore, if one is about to raise a first round, the competition is for ~$100M budget only and across all three countries (Chart 3);
- US companies dominate largest financing events, both the first and subsequent funding rounds (Chart 4). However, the average size of subsequent rounds is larger in the UK due to the exceptional round of Bllippar ($9.1M vs. $7.4M).
Industries and functions
- Horizontal, cross-industrial applications received the largest chunk of funding, but horizontal players lost their leadership compared to all vertical applications combined, which account for 57% of funding combined (Chart 5). Ecommerce / Travel / Hospitality / Leisure and Healthcare are the most well-funded verticals;
- According to wider software industry trends, horizontal players may start acquiring vertical players. For instance, Oracle has recently spent more than a billion dollars to acquire Opower, a customer engagement and energy efficiency cloud services for utilities, and Textura, construction contracts and payment management cloud services. In cognitive tech, in Q1 2016 Digital Reasoning, a horizontal cognitive computing firm, acquired Shareable, a specialized mobile clinical documentation platform;
- In terms of functions, almost half of companies is focused on core business functions as opposed to supporting functions (marketing, HR, etc.) Such companies have raised $266M or 42% of total invested capital (Chart 6);
- The largest support function targeted by startups is IT and related specialized tools, used predominantly by non-business users. The recent acquisition of Xamarin, a cross-platform development software, by Microsoft for rumored $450M indicates importance of specialized tech in a wider context.
What actually do all this cognitive technologies do?
- Modern cognitive tech goes beyond optimizing routine processes. In some industries like Finance/Insurance and Healthcare, cognitive tech companies approach tasks that typically require higher education and are hardly routine, for instance search of financial data or interpretation of medical images;
- Cognitive tech is frequently applied to analytical tasks, and tools supporting these tasks account for the largest chunk of investments (Chart 7);
- However, at the current stages investments are largely directed to tackle routine tasks or to apply cognitive tech in processes that are already dehumanized;
- Infrastructural tech is applied in research, where general AI is being developed, and in the transportation industry, which is working on creation of self-driving cars.
- Q1 2016 was a successful period for exits, representing $346M in disclosed proceeds via M&A deals and one IPO (Table 1);
- Diversity of buyers appears to be a feature of cognitive tech acquisitions. Deals sealed in Q1 2016 show that software vendors like Microsoft, SaaS leaders like Salesforce, and consumer tech giants like Apple are all interested in the sector. Moreover, cognitive tech companies themselves acquire peers;
- Notably, the UK did particularly well, despite its lower investment activity in the field compared with the US.
N.B. Data on cognitive tech investments is limited and the boundaries of the sector are not clearly defined. Therefore, estimates of investment activity may vary. All interested parties are kindly invited to discuss data shared above.
UPD: CBInsights published their view on investments activity in the wider cognitive tech. The numbers are close to my estimates, so we seem to be at the same page now – https://www.cbinsights.com/blog/artificial-intelligence-funding-trends/?utm_source=CB+Insights+Newsletter&utm_campaign=e3ee8f0a15-TuesNL_6_21_2016&utm_medium=email&utm_term=0_9dc0513989-e3ee8f0a15–87404489
Sources and notes:
 Sampling: industrial databases returned 147 cognitive tech companies that raised funding in Q1 2016, 135 were included into a long-list after data was checked and cleaned, and 123 were shortlisted as financing data were not available on 12 companies.
 Please note, here and beyond term ‘Tools’ does not represent a certain type of a product delivery, rather than a traditional definition of ‘…something (as an instrument or apparatus) used in performing an operation or necessary in the practice of a vocation or profession…’ — Merriam-Webster. Therefore a tool may be a complex platform for interpreting medical images, or a ‘simple’ app to snap and translate handwritten text on the go.
 Please note that classification of routine vs. non-routine is a subjective one, as operations that are non-routine now may be considered routine in the future.