- Wynd Network has raised $3.5 million for its inaugural product Grass.
- Grass lets users sell their unused internet bandwidth to companies training AI.
- The project says it will decentralise ownership at the end of its beta phase.
Wynd Network, a startup enabling users to monetise their unused internet bandwidth, has announced the completion of a $3.5 million seed round.
Wynd Network’s first product is Grass — a decentralised web scraping network aiming to help companies and nonprofits train AI.
Polychain Capital and Tribe Capital led the round. Polychain’s recent investments include restaking protocol EigenLayer and blockchain key management platform Cubist.
Additional funding for the round was provided by Bitscale, Big Brain, Advisors Anonymous, Typhon V, and Mozaik, among other firms.
According to Chris Nguyen, CTO of Wynd Network, Grass will produce datasets that can be traced back to their origin, making it possible to compensate people fairly for their contribution to the network.
“We’ve explored many ways to make public web data more accessible to open-source AI projects, and [we] believe that decentralisation is the only way to achieve this both ethically and efficiently,” Nguyen told DL News.
Nguyen said the funding will be used to enhance Grass’s technological infrastructure, expand its network of nodes, and refine its data verification process.
Currently, Grass lets users offer up their unused internet bandwidth through a web browser extension to earn points. The Grass website says these points will be “retroactively converted into network ownership” at the end of the product’s beta phase.
Many onlookers expect that Grass will hand over network ownership to early users through a token — similar to many decentralised autonomous organisations, or DAOs.
When DL News asked Wynd Network about a token, a representative from the company declined to comment.
According to Wynd Network, once the product leaves beta, users will be able to earn up to $30 a month selling their bandwidth.
Training AI with Grass
Training AI, such as large language models, requires the software to comb through huge amounts of data searching for patterns.
Typically, such large-scale operations are carried out from data centres with warehouses full of dedicated servers. But according to Nguyen, many websites restrict IP addresses associated with data centres, making it impossible for AI labs to use them when scraping the web.
“This is where Grass comes in,” Nguyen said. “It’s a way to access the public web from a million different IP addresses at once, at scale, in a way that is really difficult to shut out.”
Through residential IP addresses provided by users, Grass lets AI labs directly scrape the internet for web data to train their AI models.
Mitigating risks
Letting web users sell their unused internet bandwidth is not a new idea.
Such applications, commonly referred to as proxyware, have been around for years, advertising that users can generate passive income from their unused network resources.
While proxyware apps emphasise that they are safe for users, an investigation by cybersecurity software company Trend Micro showed that they can pose security risks in some cases.
Trend Micro recorded and analysed network traffic from a large number of computers using several different bandwidth sharing services.
While it found that most of the traffic was legitimate, it also identified some questionable activities, such as bulk registration of social media accounts, accessing potential click-fraud sites, and scraping government and personally identifiable information databases.
To make sure Grass’ customers don’t use its users’ bandwidth for malicious activities, the platform maintains strict standards about who qualifies to use the network.
Nguyen said that all buyers go through a rigorous vetting process, and Grass only accepts vetted organisations, such as nonprofit AI data repositories and incorporated companies that pass know-your-customer checks.
“The team has total visibility over everything buyers do on the network and can see if anything malicious is being done,” Nguyen said.
So far, the Grass app has been downloaded over 103,000 times, which is perhaps a testament to the trust users are placing in it.
The company hopes that its Android mobile app, set to release in the coming weeks, will help grow the network further.
Tim Craig is DL News’ Edinburgh-based DeFi Correspondent. Reach out with tips at tim@dlnews.com.
