Select your language: EN JP

Key Trends from Analyzing 282 AI Acquisitions in 2020

Miraj Mainali, Senior Research Associate
May 23, 2021

Lux Research has been performing quantitative analysis of the M&A activity in the AI space for the last five years. (Click here for those insights: 2016 and 2017, 2018, and 2019.)

Starting last year, we began to incorporate larger data sources, enabling us to capture a more comprehensive landscape of M&A activity. We use descriptions from publicly available M&A announcements and our own anecdotal knowledge of the space to analyze trends and predict where the AI space is headed. In this blog, we outline five key trends we found from analyzing 282 AI acquisitions in 2020.

1. M&A counts in the AI space slightly increased from 2019 despite the pandemic

Table 1: Number of AI acquisitions from 2016 to 2020

Table 1: Number of AI acquisitions from 2016 to 2020

The total number of AI acquisitions increased in 2020 from 2019 but only by a handful of companies. 2020 was an unprecedented year because of the coronavirus pandemic, which brought significant financial instability into almost every industry globally. Acquiring companies was the least of the concerns of many companies, as some had to shut down their operations, and many are still struggling from the ripple effects of those shutdowns.

As M&As are typically planned months ahead, the first quarter of 2020 had the most acquisitions (30%), which alone accounted for half the total M&As in 2019. Based on an extrapolation of this rate, there should have been more than 500 acquisitions in 2020. However, M&As in the second quarter of 2020 suddenly dropped significantly (to 19%), as it was the peak quarter of pandemic instability and shutdowns. As businesses begin to recover, M&A activity has started to pick up. The fourth quarter accounted for 28% of the acquisitions, which was a positive sign that companies had resumed their interests in AI startups as they entered 2021, with similar levels of enthusiasm as when they entered 2020.

2. The U.S.'s dominance in terms of successfully exiting AI startups remains unaffected

Fig. 1 Number of AI companies acquired in 2020 from each country

Fig. 1: Number of AI companies acquired in 2020 from each country

The U.S. was one of the most (if not the most) hard-hit countries by the coronavirus pandemic. However, although three fewer U.S. companies were acquired in 2020 than in 2019, it remains a powerhouse in cultivating successful AI startups. The U.K., Canada, and Israel continue to follow behind the U.S. It is surprising to see Switzerland's appearance in the top 6, surpassing India and France from the previous year.

Unsurprisingly, we were able to gather only two Chinese AI companies that got acquired in 2020. As we mentioned last year, our data sources might have missed acquisitions that were announced only in local Chinese languages, as they utilize press releases and news articles that are written in or translated into English.

Although we have less visibility into the Chinese AI space, the political and socio-cultural rifts between the U.S. and China, which were exacerbated in 2020, steered investors and acquirers (especially from the U.S.) away from engaging with Chinese startups. As a result, the VC funding for Chinese AI startups saw a drastic reduction from $9.6 billion in 2019 to $4.5 billion in 2020. For comparison, VC funding for AI companies in the U.S. only decreased from $25.2 billion to $23.6 billion over the same period.

3. Machine learning & AI are still most commonly used as blanket terms, while computer vision is gaining momentum

Fig. 2 Technologies used by the AI companies that were acquired in 2020

Fig. 2: Technologies used by the AI companies that were acquired in 2020

Just like last year, more than 50% of acquired AI startups vaguely describe themselves as machine learning and AI companies. Apart from that, computer vision (8.7%), deep learning (4.9%), and natural language processing (NLP) (4.5%) remain the other most common technologies used by those acquired startups. While NLP has actually regressed from the previous year's third position (after generic AI), computer vision has remained about the same.

Some of the most common use cases of computer vision in physical industries were in automated inspection, which saw increased interest from companies in 2020 due to the COVID-19 pandemic, which caused a shortage of workers due to sickness and regulations. The rapid adoption of synthetic data in 2020 was another factor for the rise of the popularity of computer vision and deep learning. On the flip side, other emerging technologies, such as transfer learning, reinforcement learning, emotion AI, and AutoML, saw decreased M&A activity in 2020.

4. While sales and marketing remain the most sought-after use cases in AI, healthcare and robotics are seeing growth

Fig. 3 Industries served by AI companies that were acquired in 2020

Fig. 3: Industries served by AI companies that were acquired in 2020

Sales and marketing, automotive, HR and recruiting, and customer support were the top use cases of AI companies in 2020. While use cases in healthcare and robotics slightly increased, other use cases in energy, gaming, pharmaceuticals, communication, logistics, cybersecurity, food and agriculture, smart home, wearables, and manufacturing didn't see much growth. This shows that companies went for proven use cases of AI in the year when they had limited resources to invest in new acquisitions. In the next two to three years, we can expect to see an increase in acquisitions of AI companies in use cases like automated inspection, drug discovery, and optimizing logistics, which have seen tailwinds from the pandemic.

5. Tech companies are no longer buying AI companies at the same rate as they have shifted toward an in-house AI strategy

Table 2 The number of AI acquisitions by big companies in 2019 and 2020

Table 2: The number of AI acquisitions by big companies in 2019 and 2020

Apple is the only GAFAM company to have acquired more companies in 2020 than in 2019. Looking at the acquisitions, Apple mostly sought out companies to improve its existing consumer products, such as Siri, podcasts, and video annotations, while also investing in emerging technologies like virtual reality (VR) for its upcoming products. As we predicted last year, these large tech companies continue to build their own AI solutions with the talent they acquired via prior acquisitions and from continuous hiring. Therefore, AI startups have become less desirable for such companies. As a result, we expect these large companies to acquire fewer startups and instead lead the pack in developing new AI applications themselves.

Lux Research Digital Deep Tech Newsletter


  • The coronavirus pandemic was a big knock to the rising demand and popularity of AI startups in 2020. However, as normalcy returns, AI startups will gain back their luster.

  • Those planning to invest in and acquire AI companies in 2021 and beyond should continue considering the U.S. and Europe as their best bet. The cultivation of AI startups in Asia remains low, and despite leading in patent filings, publications, and now citations, Chinese startups still struggle to exit through M&A successfully because of sociocultural issues.

  • Emerging technologies like transfer learning, reinforcement learning, emotion AI, and AutoML took a significantly stronger hit due to the changing market dynamics compared to popular AI techniques. Growing demands and technology enablers like drones, robotics, satellite imaging, and synthetic data are leading to the accelerated adoption of computer vision.

  • Low-hanging fruit like sales and marketing, and HR and recruiting will remain target use cases for many AI startups, as the demand doesn't show any signs of decreasing anytime soon. As many physical industries have started exploring and investing in AI, there should be a slow but steady rise in the demand for new use cases in those industries.

  • Thsose interested in acquiring AI startups should also start exploring opportunities to build AI themselves. For many companies, a hybrid of working with AI startups and internal development is still the best path forward.
Battling COVID-19 with Transformational Tech: Singapore Case Study

Battling COVID-19 with Transformational Tech: Singapore Case Study

Read More
The Digital Transformation of Supply Chain Management Report

The Digital Transformation of Supply Chain Management Report

Read More
Lux Research Establishes New Framework for Selecting Robotics Vendors

Lux Research Establishes New Framework for Selecting Robotics Vendors

Read More
More Digital Research
Schedule Your Demo