October 11, 2022

AI investment is down by 36% — the sector in numbers

Funding for artificial intelligence startups worldwide has dropped to 2020 levels

The artificial intelligence sector saw massive growth in 2021 — but investment is now back to where it was in 2020, according to the latest State of AI report.

This is mostly thanks to a drop in $100m+ "megarounds", something that other sectors have seen too.

“The reduction in AI investment is not unique to the sector — VC investments overall are down 53% year-on-year,” says Nathan Benaich, coauthor of the report and general partner at London-based VC Air Street Capital. 

Some areas of deeptech continue to pique investors’ interest though. The report highlights defence as a rapidly growing AI application — an example is Helsing, a Berlin-based AI defence startup that raised €100m in November last year. The launch of Nato’s €1bn fund last month has also given a big boost to AI companies working on defence projects.  


On the less ethically complex side of the spectrum, Benaich also thinks startups applying AI to biotechnology will boom. “I see a major inflection in AI for science and specifically companies that are addressing problems in the life sciences with AI-first approaches,” he says, pointing out examples such as British drug discovery company Exscientia and Swiss cancer diagnostics startup Hedera Dx. 

So what does the AI sector look like in 2022? Let’s look at the numbers.

Investment in AI is down

The AI sector is estimated to have raised $70.9bn by the end of 2022, according to the report. This will set worldwide AI funding back to 2020 levels after the 60% increase it saw in 2021.

The biggest culprit behind this drop is a big reduction in the amount of megarounds of over $100m. Meanwhile, smaller rounds are on track to keep up with 2021 levels. 

“The AI sector is still very healthy,” Benaich tells Sifted. “We can expect a boom in AI-first startups, particularly in creative domains, enterprise software and biotechnology.” 

AI unicorns keep multiplying

Despite the slowdown in investment, the number of unicorns in the AI sector is still growing — and fast. The US, UK, Germany, Singapore and Switzerland have more than doubled their AI unicorn count they have when compared to the numbers in last year’s State of AI report

Some notable AI unicorns in Europe are the UK’s Tractable, an insurtech startup that hit a billion-dollar valuation last year; Agile Robots, which became Germany’s first robotics unicorn in 2021; and VR startup MindMaze, which became the first Swiss unicorn back in 2016.

The report found that while public valuations of AI companies have dropped in 2022, the valuation of private companies keeps growing. “This reflects the lag in valuations being reset in private markets versus the daily reset that happens in public markets,” Benaich explains.

“As more private AI companies become funded and raise subsequent rounds, the sum of their enterprise value goes up. However, as there have been few to no IPOs this year and a significant pullback in public company valuations, the sum of their enterprise value goes down.” 

IPOs are down, acquisitions are up

The number of IPOs has dramatically dropped, from 104 in 2021 to 26 in 2022 so far.

“The bar for being 'IPO grade' has jumped dramatically due to the valuation reset — how many multiples of current and future revenue investors are willing to pay to buy shares today. This has made it more difficult for unicorns with subpar revenues to go public,” says Benaich. 

On the other hand, the number of acquisitions is on track to outperform 2021 numbers. This is because private companies are now “cheaper”. According to Benaich, private companies are now worth around 10 times their annual recurring revenue, while investors were willing to pay several times that just a few months ago.  

When looking at the EU, Switzerland and Norway, the report found that the amount of acquisitions in 2022 already exceeds last year’s figures: 97 in 2022 to date compared to 94 in 2021. 

Clara Rodríguez Fernández is Sifted’s deeptech correspondent, based in Berlin. Find her on LinkedIn.