Investors put more cash into nascent startups specializing in Web3 and DeFi than their A.I.-focused counterparts in the third quarter of last year, according to a recent report by deal-tracking firm Pitchbook. Web3 generally refers to a new era of the internet built on blockchain technologies.

“Some of the hype has departed from pure-play AI companies,” said the report, which focused on the state of A.I. investments in 2021.

The total amount of money invested in early-stage Web3 startups was six times more than in A.I.-focused ones, the report noted. An earlier report by the same company revealed that young Web3 and DeFi startups landed investments worth a combined $1.26 billion in the third quarter of 2021 while embryonic A.I. and machine-learning startups received $212.8 million.

To be sure, A.I. is still near the-top-mind for most executives and investors, with Pitchbook estimating that the entire A.I. market was worth $145.8 billion in 2021. By 2024, It’s expected to soar 92% more to nearly $280 billion. 

But the fact that venture capitalists who are focused on early-stage companies are increasingly funding Web3 startups rather than A.I. underscores a potential sea change in the tech industry. However, it’s unlikely to be felt by most businesses until years from now since these companies are so young. 

“We’re seeing a high focus from those leading investors on this particular paradigm shift on Web3 and a lesser focus on A.I.,” Pitchbook senior analyst Brendan Burke told Fortune.

It was just a few years ago that early-stage investors believed that A.I. startups would fundamentally change the technology industry. But while the A.I. industry has grown, it “might not have grown quite as quickly as expectations from the last decade,” Burke said.

Instead of creating a market worth trillions of dollars, A.I. has, so far, birthed a market worth hundreds of billions of dollars—if you can call that disappointing. 

Now, some early-stage investors believe that Web3 represents a change in how people interact with computers, Burke said. These investors believe the concept of decentralized computing, in which apps are powered by blockchain technology that lets them operate on a vast network of personal computers, represents a “primary building block” rather than a “complementary technology,” he said.

It’s unclear whether venture capitalists will be proved correct. Fundamental changes in computing, such as the rise of mobile computing and the smartphone, are difficult for many investors and companies to predict, Burke noted.

And A.I., despite being passed on by some early-stage investors, is still a hot topic with many big businesses. Early-stage investors, however, aim for the “highest growth opportunities” and are setting their sights on Web3, he said. 

Jonathan Vanian 
@JonathanVanian
jonathan.vanian@fortune.com

A.I. IN THE NEWS

Elon Musk’s Tesla robot. Tesla leader Elon Musk still plans to debut a humanoid robot sometime in 2023, despite Tesla’s A.I. director Andrej Karpathy taking a sabbatical, Fortune’s Christiaan Hetzner reported. Karpathy leads Tesla vision technology, which is key to the company’s full self-driving technology.

Sum it up for me. Google released details about a new A.I.-powered feature for Google Docs that automatically creates one-to-two sentence summaries of documents. Google researchers used a mix of A.I. technologies to create the feature, including the trendy transformer neural network software that can discover patterns in long snippets of text. Specifically, Google used an A.I. technique it dubbed Pegasus to improve the way its transformer neural network learned to summarize articles.

That sales person looks…strange. Employees at a handful of companies have created fake personas on LinkedIn to pose as salespeople trying to make first contact with potential customers, according to a NPR report. The LinkedIn profile photos were entirely A.I.-generated faces created by deepfake technology. Although most people worry about deepfakes being used for government propaganda, the LinkedIn episode shows that companies are now using the tech for less concerning, but annoying uses. From the article: Anyone who takes the bait gets connected to a real salesperson who tries to close the deal. Think telemarketing for the digital age.

Health care A.I. About 100 hospitals that are part of England’s National Health Service (NHS) network are using machine-learning technology to predict the number of patients seeking care, which helps the staff better allocate resources, tech publication The Register reported. The startup Faculty developed the A.I. software. From the article: Guessing people's ages allows the NHS to prepare freeing up beds at different departments to care for children at pediatric units or provide better support for elderly patients when rates of A&E admission are expected to be high.

EYE ON A.I. TALENT

Cowbell Cyber has hired Joshua Chan as the cybersecurity insurance company’s chief technology officer. Chan was previously the vice president of engineering at security firm Zimperium.

The City of Philadelphia promoted Sara Hall to be its director of digital services, the local tech news site technical.ly reported. Hall was originally a product manager when she began working for the city. 

EYE ON A.I. RESEARCH

A.I. for the ecological impact. Researchers from institutions like the University of Gothenburg and the University of Fribourg published a paper in Nature Sustainability about using the A.I. technique of reinforcement learning—in which computers learn by numerous trials—to help calibrate the optimal biodiversity settings to improve ecological conservation. The A.I. system can more accurately model how logging could impact the ecological conditions of a forest than existing methods, the researchers wrote.

From the paper: Within AI, we implemented a reinforcement learning (RL) framework based on a spatially explicit simulation of biodiversity and its evolution through time in response to anthropogenic pressure and climate change. The RL algorithm is designed to find an optimal balance between data generation (learning from the current state of a system, also termed ‘exploration’) and action (called ‘exploitation’, the effect of which is quantified by the outcome, also termed ‘reward’).

FORTUNE ON A.I.

Invading Ukraine has upended Russia’s A.I. ambitions—and not even China may be able to help—By Jonathan Vanian

Companies are turning to A.I., but C-suite collaboration is crucial for success—By Sheryl Estrada

These investors went through IVF. Now they’re putting $22 million behind an A.I. fertility startup—By Emma Hinchliffe

Amazon’s Alexa is becoming a member of the family—By Stephanie Cain

Apple and Google criticize the new EU Digital Markets Act that will radically change the way they have operated for the past 20 years—By Sophie Mellor

The miseducation of A.I.—By Jonathan Vanian

BRAIN FOOD

When A.I. drones strike Ukraine. Eye on A.I.’s Jeremy Kahn explored the use of A.I.-powered military drones in Russia's invasion of Ukraine. Newer semiautonomous drones are smaller and cost less than hulking Predator drones, making them attractive weapons for militaries. These drones also require less training of soldiers before they can deploy the weapons in the field. Kahn wrote that “companies building loitering munitions and other A.I.-enabled weapons say they are meant to enhance human capabilities on the battlefield, not replace them.” But critics of more advanced autonomous weapons are concerned about their growing use.

From the article: Noel Sharkey, emeritus professor of computer science at the University of Sheffield in the U.K., who is also a spokesperson for the Stop Killer Robots campaign, says automation is speeding the pace of battle to the point that humans can’t respond effectively without A.I. helping them identify targets. And inevitably one A.I. innovation is driving the demand for more, in a sort of arms race. Says Sharkey, “Therein lies the path to a massive humanitarian disaster.”