Based on analysis on our AI research platform, we see that exit activity in the first half of 2018 is slightly down from 2017.
2018 Mid-Year AI Exit Activity Lower Than 2017 But Higher Than 2016
Let’s take a closer look at the number of AI exit events by year.
The above graphic shows 32 exits in the first half of 2018. For the past three years, Q3 and Q4 accounted for 46% of total exit events on average. If that trend holds, 2018 exits finish the year slightly lower than 2017, but higher than 2016. We’ll see if the second half of the year changes this trend!
Last quarter we observed that the artificial intelligence sector is maturing. This quarter we are conducting a deeper analysis on our AI research platform to examine funding by category. Our analysis shows two important observations:
Machine Learning Platforms and Computer Vision Platforms lead the sector in Q2 funding
Machine Learning Applications dominates the sector in all-time funding
We’ll explain these observations with some graphics and discussions below.
Machine Learning Platforms and Computer Vision Platforms Lead AI in Q2 Funding
To start off, let’s scrutinize the AI funding by category in Q2.
The above graphic shows that both Machine Learning Platforms and Computer Vision Platforms lead the sector in Q2 funding with $1.5B each. Machine Learning Applications and Smart Robots follow in the second and third places with $1.4B and $1B, respectively. It’s also noteworthy that there is a steep drop-off after Smart Robots, as its funding is 3.4 times higher than the next category, Speech Recognition.
So we’ve witnessed how different AI categories stack up in their Q2 funding. But how do these categories’ funding compare with each other historically? Let’s investigate that in the next section.
Machine Learning Applications Dominates AI in All-Time Funding
The graph below shows the all-time funding for different AI categories. The quarterly funding and growth rates of these categories are also highlighted.
The bar graph indicates Machine Learning Applications completely dominates the sector with $17B in total funding. This is more than twice the funding in the next category, Machine Learning Platforms.
In addition, the line graph demonstrates that Computer Vision Platforms saw the highest growth rate in Q2 at 48%.
Conclusion: Machine Learning Categories Are At the Forefront of AI Funding
In summary, we have analyzed the AI funding amounts in different categories. We’ve discovered that Machine Learning Platforms and Computer Vision Platforms lead the sector in Q2 funding. In addition, Machine Learning Applications dominates AI in all-time funding. It’ll be interesting to see if any other AI categories will catch up in the rest of 2018.
The Deep Learning Applications category leads in the number of exits
The Computer Vision Platforms category leads in acquisition amount
We’ll illustrate these takeaways with some graphics that show AI exit activity by category.
To help set the stage, the graphic below shows AI exit activity over time. As you can see, the sector’s exit activity experienced strong growth over the past few years.
Deep Learning Applications Leads AI in the Number of Exits
Let’s examine the exit events for each AI category. Exit events include both acquisitions and IPOs. The below graph highlights the number of AI exit events by category.
This graph shows that the Deep Learning Applications category leads the sector with 71 exit events. Natural Language Processing comes next with 46 exit events.
Deep Learning Applications includes companies that utilize computer algorithms to optimize operations in vertically specific use cases. Examples include using deep learning technology to detect banking fraud or to identify relevant sales leads. Some example companies are Sift Science, SparkCognition, Sumo Logic, and BenevolentAI.
Let’s now see how AI categories compare with one another by acquisition amount.
Computer Vision Platforms Leads AI in Acquisition Amount
The graph below shows the acquisition amounts in different AI categories.
We can see from this graph that the Computer Vision Platforms category leads all the other AI categories by far. The total acquisition amount in this category is around $16 billion. Computer Vision Platform companies process images to algorithmically derive information from them and recognize objects. Some example companies in this category include Cortica, Blippar, Kairos, and Clarifai.
Computer Vision Platforms has seen some large acquisitions in recent years. Mobileye was acquired by Intel in March 2017 for around $15 billion. Movidius was acquired by Intel in September 2016 for $400 million. Magic Pony Technology was acquired by Twitter in June 2016 for $150 million.
The acquisition amount in Computer Vision Platforms represents 72% of all AI acquisition activity. It’s noteworthy that its acquisition amount is more than ten times the next category, Deep Learning Platforms. Additionally, Computer Vision Platforms’ acquisition amount is highly concentrated, with 15/16 of the amount coming from the $15 billion Mobileye acquisition.
Conclusion: Deep Learning Applications and Computer Vision Platforms Lead AI Exit Activity
In summary, we have examined AI exit activity by the number of exit events and acquisition amount. The Deep Learning Applications category leads the sector in the number of exit events. The Computer Vision Platforms category leads in acquisition amount. It will be interesting to see which other categories take the lead in AI exit activity in the rest of 2018.
We will now examine the different components of AI and how they make up this startup ecosystem. On our AI research platform, we have classified the companies into 13 categories. This blog post illustrates what these categories are and which categories have the most companies. We will also look at how these categories compare with one another in terms of their funding and maturity.
Machine Learning Applications Is the Largest AI Category
Let’s start off by looking at the Sector Map for the AI sector. As of March 2018, we have classified 2161 AI startups into 13 categories that have raised $32 billion. The Sector Map highlights the number of companies in each category. It also shows a random sampling of companies in each category.
We see that Machine Learning Applications is the largest category with 762 companies. This category contains companies utilizing algorithms that learn and optimize automatically from data. These companies tackle issues in specific use cases such as detecting banking fraud or identifying top retail leads. Some example companies include Sift Science, SparkCognition, Sumo Logic, and BenevolentAI.
We have seen what the different categories making up this sector are and the number of companies in each. What about their funding and maturity in relation to one another? Let’s look at our Innovation Quadrant to find out.
Most of the AI Categories Are Pioneers
Our Innovation Quadrant divides the AI categories into four different quadrants.
We see that the Pioneers quadrant has the most AI categories with 8. The Pioneer categories are in the earlier stages of funding and maturity. Speech Translation and Speech Recognition landed in the Established quadrant. Both of these speech technology-related categories have reached maturity yet with less financing.
Smart Robots, Recommendation Engines, and Machine Learning Platforms are in the Disruptors quadrant for acquiring significant financings at a young age.
We’ve now seen the AI categories and their relative stages of innovation. How do these categories stack up against one another? Let’s look at the Total Funding and Company Count Graph.
Machine Learning Applications Startups Have the Most Funding and Companies
The graph below shows the total amount of venture funding and company count in each category.
As we’ve seen in the Sector Map and the above graphic as well, the Machine Learning Applications category leads the AI sector with 762 companies. The above graphic also shows that Machine Learning Applications leads in funding with almost $16 billion. Some of the best-funded companies in this category include Toutiao ($3B), Argo AI ($1B), Indigo ($359M), and Wecash ($328M).
The funding in Machine Learning Applications is more than 267% of that in the next category, Machine Learning Platforms. These two categories are related yet have different functions. Machine Learning Applications companies apply self-learning algorithms to optimize specific business operations. Machine Learning Platforms companies build these self-learning algorithms or their underlying infrastructure.
Conclusion: Machine Learning Applications Category Dominates the AI Sector
From the above analysis, we can see that Machine Learning Applications dominates the AI sector in total funding and company count. It’ll be interesting to see how the AI landscape will change and develop throughout the rest of 2018.
What are your thoughts on this? Let us know in the comments section below.
Funding amounts are shifting to mid and late-stage events
Later-stage funding counts increased while early-stage funding counts dropped
We’ll explain these takeaways with some graphics that show AI funding activity by round.
To help set the stage, the graphic below illustrates AI funding amount over time. As you can see, the sector’s funding exploded in 2017.
AI Funding Amounts Shifting to Mid and Late-Stage Events
We’ll start off by examining the annual AI funding amounts. The below graph shows recent AI funding amounts in different rounds.
We see that AI funding amount for all stages grew from 2012 to 2017. Most notably, Series D funding grew the most.
Let’s look at the AI funding amount by round as a percentage, which shows changes independent of the total size.
From 2012 to 2017, Seed and Series A funding amount percentages dropped. During the same period, Series B, C, and D funding amount percentages increased.
Specifically, the Seed round funding amount dropped from 15% to 3% from 2012 to 2017. The Series A funding amount dropped from 40% to 20%. On the other hand, the Series B funding amount increased from 26% to 30%. The Series C funding amount increased from 15% to 25%. The Series D funding amount increased from 4% to 20%.
Taken together, these two graphics show that the AI funding amounts in Series B, C, and D increased from 2012 to 2017. The funding amount as a percentage graph also shows that the Seed and Series A funding dropped.
So we have seen the funding amounts graphs indicate a shift to mid and late-stage events. Would the funding event count graphs show the same trend? Let’s examine them in the next section to find out.
Later-Stage AI Funding Counts Grew; Early-Stage Funding Counts Dropped
Let’s now look at the annual AI funding event counts. The below graph shows the AI funding counts in different rounds over recent years.
The above graph shows that Seed round events grew from 93 in 2012 to its peak at 275 in 2015. It then fell from 275 events to 195 events in 2017. On the other hand, the funding event counts for all the other stages grew from 2012 to 2017.
Specifically, Series A funding events grew from 40 to 176. Series B funding events grew from 12 to 84. Series C funding events grew from 4 to 35. Series D funding events grew from 2 to 21. Late Stage funding events grew from 0 to 9.
Let’s now look at the AI funding count by round as a percentage, which can show shifts more clearly.
This graph shows that from 2012-2017, the Seed round funding count percentage dropped. In the same years, the funding count percentages for all the other stages increased.
Specifically, the Seed round funding count dropped from 62% to 38% from 2012 to 2017. In contrast, the Series A funding count increased from 26% to 34%. The Series B funding count increased from 8% to 15%. The Series C funding count increased from 3% to 6%. The Series D funding count increased from 2% to 4%. And the Late Stage funding count increased from 0% to 2%.
Combining these two graphics, we can see that the AI funding count for the Seed round dropped from 2012 to 2017. The funding counts for all the other stages grew by various degrees during the same period.
Conclusion: The AI Sector Is Maturing
In conclusion, we have seen that AI funding amounts have shifted to mid and late-stage events. AI funding counts also saw an increase in mid and late-stage events and a drop in Seed events. These observations led us to conclude that the AI sector is maturing. Investors seem to be placing heavier bets into more well-established AI companies.
What are your thoughts on this? Let us know in the comments section below.