AI Startups You Should Know – Part 1

In this article, I want to showcase some disruptive AI startups by application. My first article outlined the overall fundraising trends across the global AI landscape. My second article took a closer look at how conglomerates invest/acquire AI startups. Now, I hope to shine light on actual use cases of AI and what startups you should be paying attention to.

I mentioned before that I reviewed 1,600+ startups globally. After looking through each company, I was able to split most startups into two main categories (Horizontal and Vertical), of which was split into 22 subcategories to better define their application.

In this article, I only want to focus on the Horizontal AI startup category. This is made up of 9 subcategories, which are Analytics, Assistant, Data, Infrastructure, NLP, Robotics, Security, Tool and Vision (The definition of each subcategory is outlined in my previous article). I’ll do my best of giving a general overview of the subcategory and then I will mention a few startups that are excelling in this subcategory.

 

Analytics: (AI built solely to analyze general data and provide insight)

The Analytics AI subcategory is challenging because it’s hard to distinguish startups that are building a glorified business intelligence tool and an analytics application that has a form of machine intelligence. Most startups claim they can provide “transformative” information based on existing big data, but the trends that I’m seeing are Analytic AI startups are building tools that can automatically identify patterns in big data sets. This is different from most BI tools that require users to understand what they’re looking for or ask questions to probe for patterns. These tools simply figure out all possible combinations and suggests patterns/factors you may not have noticed.

Most of these Analytic AI startups cater to enterprises trying to optimise operations, understand customer behaviour, manage vendors/finances, etc. Outside of enterprises, this subcategory uses AI for a number of analytical purposes. Examples include natural disaster damage estimates, predicting maintenance of machinery, discover patterns/information across social networks, optimise mobile networks, identify opportunities in financial markets, identify economic or political changes, predicting price changes in consumer products or airfare, analyse scientific or legal literature, helping realtors figure out when someone is moving, predicting success of books or movie scripts before release and much more.

Startups in this subcategory you should watch include: Ayasdi*, ThoughtSpot, Tamr, INCORTA, Banjo, FLYR* and Interana*.

In total are about 20 out of 100+ AI startups in this subcategory that I’m watching closely.

 

Assistant: (Tools that are built to be forms of chat bots)

The Assistant subcategory is comprised of mostly startups that create chat bots. Assistant AI startups can take many form factors but their ultimate goal is to help you complete a task through a simply question/answer system.

Applications of chat bots are vast but each chat bot is only capable of handling a single use case without human intervention. Applications include setting up meetings, querying information from science/financial/legal databases, managing customer service questions, travel/flight/hotel assistant, personal on-demand assistant, virtual sales teams, banking requests, buying gifts and more.

It’s notable that Tech Giants were very active in this category between 2013-2015, but interest dropped off afterwards.

Startups in this subcategory you should watch include: Eloquent Labs, Octane AI, Hyper Anna, X.ai, Mezi (Recently Acquired), Magic*, Recast.ai (Recently Acquired), Action.ai and Sensay*.

In total are about 10 out of 60+ AI startups in this subcategory that I’m watching closely.

 

Data: (Startups that acquire, analyze and sell private data sets)

The Data subcategory is very similar to that of the Analytic subcategory. The only difference is that these startups hold proprietary data. Many of these startups offer services to help you cross reference their data with yours, others provide data for educational purposes, while most simply sell their data/analysis to give companies an edge over their competitors.

Startups in this subcategory you should watch include: Near*, Qloo*, Node*, Premise*, and Descartes Labs.

There are 5 AI startups out of 15 that I’m watching in this subcategory.

 

Infrastructure: (Startups building the physical or software infrastructure to empower AI tools)

The Infrastructure subcategory is comprised of mostly hardware AI startups. These companies are focused on building servers, chips or any hardware system that can process information faster, handle more complex calculations and/or can operate more efficiently. There is a blend of quantum computing in this subcategory. Almost all of these startups are focused on solving the infrastructure challenges required for AI tools to perform optimally.

Startups in this subcategory you should watch include: Graphcore, Kneron, Cambircon Technologies, Rigetti Computing*, Thinkforce, LeapMind, Bigstream* and Wave Computing*.

There are 10 AI startups out of 30 that I’m watching in this subcategory.

 

NLP: (AI startups focused on understanding speech, voice or text)

The NLP subcategory is made up of startups building tools that pick up on human voice/text, understand the meaning behind what is said/written, and sometimes take action upon that information. What differentiates most NLP’s is how they compare your voice/text to a particular data set. Applications of NLP include translation, interacting with electronic devices, querying a knowledge base of an organisation, analysing or optimising meetings/sales/customer service and more.

Startups in this subcategory you should watch include: MindMeld (Acquired), Unbabel, SoundAI, AISense, Insight Engines*, Entefy*, Snips*, Agolo* and Semantic Machines*.

There are 11 AI startups out of 60+ that I’m watching in this subcategory.

 

Robotics: (AI built for robots to perform any types of tasks, movements, etc)

The Robotics subcategory is broken in to two parts. First, there are startups focused on the ‘brains’ of a robot. Their goal is to train robots to perform various tasks on their own. This is done training the robot to teach itself or users guiding the robot to perform a specific task. Second, there are startups building robots solely for a specific task.

Robots are built to be companions, some for healthcare monitoring, managing warehouses or retail stores, security guards, inspection, delivery, restaurant catering, education and more.

Startups in this subcategory you should watch include: Vicarious*, Intuition Robotics*, CloudMinds*, Embodied*, ROBART GmbH and Embodied Intelligence.

There are 14 AI startups out of 45+ that I’m watching in this subcategory.

 

Security: (Tools built to protect companies from security threats)

Security is a challenging subject when reviewing startups because it is very arbitrary how startups measure success. Most Security AI startups claim they are better a spotting abnormalities. How security startups take action upon their findings are different. Some simply notify companies of patterns that aren’t normal, others provide action for specific cases and others try to solve the problem.

Examples of applications in this industry include identifying malware in emails/computers/networks, monitoring phones/IoT/Vehicles/enterprises, various forms of authentication, identifying fraud in various financial services, monitoring your online presence or your online community and more.

Startups in this subcategory you should watch include: Emailage, Versive, UnifyID*, Callsign*, Simility, NS8 Inc, Blue Hexagon, ZingBox and Socure

There are 25 AI startups out of 60+ that I’m watching in this subcategory.

 

Tool: (Startups building AI algorithms)

The Tool AI subcategory focused on building tools to optimise, deploy and manage machine learning models. Tool AI is mostly built for developers but there are many startups trying to help non-programmers deploy machine learning models. Many of these startups offer different analytical services based on the common trends they find among their clients. These services are very similar to that of Analytics AI subcategory. Many of these startups offer a consulting services to deploy their tools.

This subcategory is popular among investors from all industries. In addition, there a handful of startups that have multiple investors from the same industry.

Startups in this subcategory you should watch include: Kensho*, Bonsai*, H2O.ai*, CognitiveScale*, DataRobot*, Element AI* and Neura*.

There are 30+ AI startups out of 95+ that I’m watching in this subcategory.

 

Vision: (Startups building image recognition tools)

Vision AI startups are applying forms of image recognition to various applications. Examples include indexing content of videos, visual search engines, analytics of earth imagery, security and monitoring of offices or cities, identifying people or understanding their emotions, capturing key moments of an event, building eyes for autonomous vehicles/drones/objects, home designing, analysing safety in construction sites, building logos and much more.

As you may recall from my previous article, this was the most heavily invested category by large corporations. Tech Giants, Top VCs and Commerce were the most active in this subcategory.

Startups in this subcategory you should watch include: Mighty AI*, PointGrab Ltd.*, Vidrovr, Clarifai*, PrecisionHawk*, Skycatch*, Shield AI*, Vion Technology and SenseTime

There are 40+ AI startups out of 130+ that I’m watching in this subcategory.

 

Conclusion:

I hope this review gives you more clarity about what is actually happening in the AI ecosystem. My goal was to share different examples of AI in order to make the subject more familiar. Ultimately, I hope this serves to inspire new founders, help corporations discover new applications of AI and help investors find new opportunities. I will publish an article soon on the Vertical AI category.

* Is for investors to know which AI startups from this list will be fundraising again in the near future, are currently fundraising for their next round or they could announce the closure of a new round very soon.

How Corporate Giants Invest In AI

Last week I published an article on the overall trends in the AI industry. After reviewing 1600+ AI startups globally, I was able to group most startups into 22 different subcategories, reveal fundraising trends, identify countries leading the AI race and more.

Taking it one step further, I wanted to share how the largest corporations by industry invest in AI startups. Watching corporate venture capital fund’s activity is a key indicator in spotting industry disruption. Corporate funds typically have very rigorous investment criteria and chase after investments that either add new revenue streams, boost operational efficiencies or attempt to partner with future competition.

I analysed over 200 of the most active CVC funds as well as the most influential corporations across 11 industries. My goal was to find which AI subcategories receive the most funding by industry, what corporations lead in AI investment per industry, what are the most popular AI startups in each industry and what AI startups are making the most impact across all industries.

The 11 industries I researched are as follows:

  • Tech Giants (Top 15 largest or most influential technology corporations)
  • Finance (Largest most active banks or financial institutions globally)
  • Telecoms (Largest telecoms by continent)
  • Electronic Manufacturers (The largest semi-conductor, chip and electronic device manufacturers)
  • Media (Largest media, advertising or entertainment corporations)
  • Industrial (Largest industrial manufacturing corporations that includes construction, appliances, metals, aerospace, chemicals, etc)
  • Commerce (Worlds largest eCommerce and retail stores)
  • Insurance (Largest life, health, automobile or reinsurance providers)
  • Consulting (The largest strategy, tech and accounting consulting firms)
  • Automotive (The largest automotive or parts manufacturers)
  • Healthcare (The largest hospitals, biochemical or pharmaceutical corporations)

I also took TechCrunch’s 2017 top 10 VC list and used it as a reference point to compare against corporations. VC’s typically have a much longer term vision than corporations and should indicate what AI startups will have more of an impact in 5 to 10 years.

Please note, all of this information is what I was able to find publicly. I’m sure there are many undisclosed investments made by these corporations. However, I think this subset of information will give you a general idea of the common trends among industries.

 

AI Investment Analytics By Industry

There was over 400 investments/acquisitions made by large corporations in AI startups. The Top VC’s and Tech Giants outpace all other industries with the most investments in AI startups. However, it’s surprising to find Finance, Telecom and Media industries are leading outside of the tech industry. Surprisingly, the Healthcare industry invested in the least amount of AI startups.

Most industry giants are investing in more US AI startups than INT startups. This means corporations abroad are pushing more capital to the US and less US investors are moving capital outside of the US. In reference to my previous article, given that there are significantly more AI startups built in the US, this is probably why we have this disparity. It is surprising however, to see the Top VCs actually invest less abroad than most tech giants.

In my previous article I split all AI startups into two categories. First by Horizontal, which are startups building AI tools, think of this as the hammer or measuring tape. Second by Vertical, which are startups creating a service using a form of AI, think of this as the plumber or construction worker. The majority of all industry leaders are investing in slightly more Horizontal AI categories than the Vertical AI categories. This is probably driven by corporations desire to find tools that improve existing products and services as opposed to trying to create a whole new product offering.

AI startups that fundraise from corporations are raising significantly more capital than the global benchmark of $10-11M. Startups receiving capital from corporations signals to the industry that there are potential synergies, therefore future growth expectations are much higher and price goes up. International AI startups raise a significant amount more capital than US AI startups. This can be a result of a tremendous amount of capital abroad with a very little supply of AI startups which inevitably drives up prices.

If you were to break apart the average fundraise by location in to categories it will paint a slightly different picture. It seems US corporate investors either find more opportunities/impactful AI startups in the Vertical category or less supply of them in comparison to the Horizontal category. For the international community it is opposite that of the US.

There doesn’t seem to be any outliers by country that pushed the average capital raise, which is something we saw in the previous article. The countries that receive the most investment from each industry and Top VCs are the United Kingdom, China, Israel, Canada, Japan and Singapore. These countries lead because they have very active corporations investing in their community and they also lead with the most amount of startups/funding outside the US.

 

AI Subcategory Investment Analytics By Industry

The graph above showcases how corporations from each industry are investing across the 9 Horizontal AI subcategories. In my previous article, most capital flowed into Vision, Tool and Analytics subcategories, this follows the corporate industry trends as well. To help you parse through the graph above, I broke down the top subcategories per industry:

  • Tech Giants: Vision, Tool and Analytics
  • Top VCs: Vision, Analytics and Security/Assistants
  • Finance: Tool, Security and Vision
  • Telecom: Vision, Analytics, Security and Assistants
  • Electronic: Tool, Security and Infrastructure
  • Media: Vision, Analytics and Data
  • Industrial: Vision and Tool/Analytics/Infrastructure/Robotics
  • Commerce: Vision and Infrastructure
  • Insurance: Tool and Analytics
  • Consulting: Security
  • Automotive: Robotics
  • Healthcare: Vision

This graph represents how corporations across all industries have invested in the 13 Vertical AI subcategories. In my previous article, Automotive, Healthcare and Fintech subcategories received the most funding. The graph above showcases how Fintech, Enterprise, CRM and Marketing are the most popular subcategories among all industries and Top VCs. This means corporations are focused on investing in operational efficiencies and customer acquisition. However, if you were to remove Top VCs, the most popular subcategory is Automotive, followed by Marketing, CRM, Fintech, Enterprise and IoT. In addition, Top VCs invest significantly more in Healthcare than they do in the Automotive subcategory. The top subcategories per industry are as follows:

  • Tech Giants: CRM, Marketing and IoT
  • Top VCs: Enterprise, Healthcare and Fintech
  • Finance: Fintech
  • Telecom: Enterprise and Automotive
  • Electronic: Automotive
  • Media: Marketing
  • Industrial: IoT
  • Commerce: Commerce
  • Insurance: Fintech
  • Consulting: CRM
  • Automotive: Automotive
  • Healthcare: Enterprise and Healthcare

 

Top AI Investors And Startups By Industry

To recap some of the information above, I wanted to share who are the most active AI investors in each industry, the most popular AI startups in some industries, and the most popular AI startups overall. The top investors and startups are below:

  • Tech Giants: Most active investors are Intel, Google and Salesforce. The most popular AI startups are MindMeld, CognitiveScale and Unbabel.
  • Top VCs: Most active investors are NEA, A16Z and Khosla/Accel. The most popular AI startups are Kensho and Timeful.
  • Finance: Most active investors are Bloomberg, Mastercard and Goldman/Fidelity/Citi. The most popular AI startups are Kensho, Moneytree and H2O.ai/Appzen/Versive.
  • Telecom: Most active investors are Softbank, NTT and Singtel. The most popular AI startups are Precision Hawk and MindMeld.
  • Electronic: Most active investors are Dell and Nvidia.
  • Media: Most active investors are R/GA, Comcast and KBS. The most popular AI startups are Vidrovr and Indicative.
  • Industrial: Most active investors are GE, ABB and Bosch. The most popular AI startup is Maana.
  • Commerce: Most active investor is Alibaba. The most popular AI startup is SalesPredict.
  • Insurance: Most active investor is New York Life.
  • Consulting: Most active investor is Bain.
  • Automotive: Most active investors are Ford and Toyota.
  • Healthcare: There weren’t any investors that stood out as the most active.

The startups with funding from the most amount of industries are CYNGN, Lemonade, Graphcore, Vicarious, ABEJA, Tamr and MindMeld.

The startups with the most corporate investors are Kensho, CYNGN, MindMeld, Moneytree, Maana, Lemonade, H2O.ai and Bonsai.

 

Conclusion

I hope this gives you a better overview of the AI startup ecosystem. As a reminder, this only showcases public information corporations are willing to share about their investments in AI. However, I think it is a good indicator where AI is growing, where it is not, and where it could be in the future.

I will continue to post a few more articles on the state of the AI industry. This will include top startups by subcategory. Stay tuned!

Kyle

Who Is Leading The AI Race?

Artificial Intelligence is becoming an extremely important function of the next evolution of technology. Therefore, I looked at every AI startup at the end of 2017. This was a collection of 1600+ startups globally. My goal was to understand the applications of AI/ML in every industry, the current fundraising environment and what countries are leading in AI development.

My summarised findings are below. You will see what countries are leading in the AI development and fundraising. In addition, I broke apart all 1600+ startups into 22 subcategories by function and application. This will help you understand where AI leads in development, funding and exits for each industry and application.

This report will give entrepreneurs, investors and companies a sense of where opportunities in AI exist and a benchmark for valuations.

For reference, I only looked at startups that have raised Angel-Series C funding. I combed through every company and applied my best judgement to keep startups if they actual apply a form of AI.

 

Countries Leading AI Development

The United States generated almost 1.5x the amount of startups than all other countries outside the US combined (International). It’s intriguing to see the United Kingdom as the leader of AI development outside the US, with Canada and Israel not too far behind.

Countries leading by total investment into AI startups paints a different picture. United States generated almost 2x the amount of investment than the entire international community received. However, China, Hong Kong and Japan have pushed a significant amount of investment into a smaller number of companies.

For both the US and all International companies, the average funding for AI startups hovers around $10M-$11M. However you’ll find that Hong Kong, China and Japan startup’s average raise is much higher. This could imply these startups are overvalued or AI startups in these markets are performing very well. Without these outliers, the average funding for International AI startups is closer to $5M.

It’s interesting to note the US as 3x as many more exits than the International community. Runner up is actually Israel, United Kingdom, Canada and China.

 

AI Startup Subcategories

After looking at over 1600 companies, I was able to come up with 2 main categories for AI startups, that further breaks into 22 subcategories. The two main categories are Horizontal AI and Vertical AI.

The Horizontal AI category are startups that are building AI tools, platforms or infrastructure that can be used across almost all industries. Think of this category as the hammer, wood or measuring tape of AI. This category breaks into 9 subcategories outlined below:

  • Analytics: AI built for analysing and pulling insights from large data sets
  • Assistant: Tools that are built to be forms of chat bots
  • Data: Startups that acquire, analyse and sell private data sets
  • Infrastructure: Startups building the physical or software infrastructure to empower AI tools
  • NLP: AI startups focused on understanding speech, voice or text (natural language processing)
  • Robotics: AI built for robots to perform any types of tasks, movements, etc
  • Security: Tools built to protect companies from security threats
  • Tool: Startups building general AI algorithms
  • Vision: Startups building image recognition tools

The Vertical AI category is applying AI tools to a particular industry to provide a service. Think of these as the plumber, construction worker or appraiser – they are simply using tools from Horizontal AI to provide unique services. This category is broken into 13 subcategories outlined below:

  • Agriculture: Startups building AI services to improve farming
  • Automotive: Startups building navigation, self-driving, maintenance, or any applications for vehicles
  • Commerce: AI companies built for different forms of commerce
  • Consumer: AI startups built for consumer lifestyle
  • CRM: AI startups focused on sales or managing customer questions
  • Education: AI startups focused on bettering education for students, teachers or schools
  • Enterprise: AI startups solely focused on improving operations in corporations, including accessing/sharing information, automating menial tasks, etc
  • Fintech: AI tools built for banking, investment, wealth management or efficient/safer ways to transact with one another
  • Gaming: Tools build for better gaming experiences, content or distribution
  • Healthcare: Any forms of managing, monitoring or servicing healthcare
  • IoT: Tools that collect, analyse or manage IoT device information
  • Marketing: Different forms of advertising or marketing activities
  • Recruiting: Tools built to source, analyse and track potential hires and or existing employees

 

Horizontal And Vertical Subcategory Break Down

Both the US and International community have built more Vertical AI startups than Horizontal AI startups.

However, the International community actually raises more funding for the Horizontal AI category than their Vertical AI category.

 

Horizontal AI Subcategory Statistics

In Horizontal AI, most US startups are created in the Analytics, Vision, Tool and Security subcategories. The International community differs slightly being that Vision, Analytics, Tool and NLP subcategories lead.

In the Horizontal AI Category, more funding is pushed into Tool, Analytics, Robotics and Vision subcategories for US startups. Whereas International startups it is Vision, Robotics, Infrastructure and Tool subcategories. This can be slightly skewed by the China, Hong Kong and Japan funding environments.

In the Horizontal AI category, you’ll notice most subcategories are at or above the average funding rate for all AI startups. You’ll also notice the Infrastructure subcategory gets a substantial amount of funding outside the US. This anomaly is from the China, Hong Kong and Japan funding environment.

It’s notable that the Analytics subcategory leads in the number of exits compared to all other subcategories in US and International.

Outside the US, the UK, Israel, China, Canada, and France are leading the development of Horizontal AI startup subcategory.

 

Vertical AI Subcategory Statistics

In the Vertical AI Category, most US startups are created in the Marketing, CRM, Enterprise and Healthcare subcategories. Whereas in International it is Fintech, Enterprise, Healthcare and Marketing subcategories.

In the Vertical AI category, more funding is pushed into Automotive, Healthcare, Fintech and Marketing subcategories. Whereas in International it is Healthcare, Fintech, Marketing and CRM subcategories.

In the Vertical AI category, you’ll notice US startups in Automotive and Agriculture receive a substantial amount of funding compared to the industry average. Outside the US you’ll find most subcategories raise below the average funding rate for AI startups.

Most exits in Vertical AI are within Marketing, CRM or Enterprise subcategories. However, it is surprising that the Consumer subcategory is one of the top exit subcategories both in US and International.

Outside the US, the UK, Canada, Israel, Germany and India are leading the development of Vertical AI startup subcategory.

 

Conclusion

There are many inferences you can pull from this data. I hope this helps puts the AI industry into perspective for many investors, entrepreneurs and corporations. It’s clear the United States outpaces all other countries in AI development and funding. Nonetheless, capital is flowing into this industry quickly and there are many opportunities to fill in each subcategory globally.

Of the 1600 AI startups, I believe about 360 will have big impact on their subcategory. Please connect or follow me as I will break down what companies to watch for in each subcategory in the future.