Multicoin Capital Backs Grass as the DePIN Data Network Pivots Toward AI Search

Multicoin Capital, one of the more prominent crypto-focused venture firms, has announced an investment in Grass, a decentralized physical infrastructure (DePIN) project that pays people to share their unused internet bandwidth and turns that capacity into web data for artificial intelligence companies. The GRASS token rallied sharply on the news.

What was announced

Multicoin general partner Shayon Sengupta said both the firm’s hedge fund and its venture fund have taken positions in Grass. The size of the investment was not disclosed. The announcement landed on September 30 and was reported by outlets including KuCoin News, PANews and Startup Fortune.

Grass has raised from well-known crypto investors before. According to Startup Fortune, it closed a $3.5 million seed round in 2023 led by Polychain Capital and Tribe Capital, followed by a Series A in September 2024 led by Hack VC.

How Grass works

The idea is straightforward. Participants install software that shares a slice of their idle home internet connection. Grass routes requests for public web pages through that residential network, parses the results and packages the data for AI developers who need large volumes of fresh, real-world web content to train and run their models.

Contributors are rewarded for the bandwidth they supply. The project says more than 6 million people now participate, which would make it one of the larger consumer-facing DePIN networks.

The revenue numbers

What makes this announcement stand out in a sector often criticized for token incentives with little real demand behind them is that Grass is publishing revenue figures. Reports citing the project say it generated roughly $17 million in the first half of 2026, with a similar figure reported for an earlier period, and that nearly all of its AI customers renewed their contracts. Grass says it is already profitable, with monthly operating costs reported at around $2 million to $3 million.

The project has also put out a forecast of about $75 million in training-data revenue for 2026. That is guidance, not a result, and it implies a steep ramp in the second half of the year. Readers should treat it with appropriate caution until actual numbers are reported.

The pivot: from training data to AI search

The more interesting part of the story may be where Grass is heading. Beyond selling bulk pre-training data, the project has launched Content API and Search API products designed to let AI agents fetch live web pages in real time through its residential network rather than through conventional scraping infrastructure.

The pitch is that AI assistants and agents increasingly need current information at the moment a user asks a question, not just a static training set. If that demand keeps growing, a network that can retrieve fresh pages at scale could address a much bigger market than training data alone. Multicoin’s investment thesis, as described in coverage, leans heavily on that search and inference angle.

Where it fits in the DePIN landscape

DePIN projects try to crowdsource real-world infrastructure, from wireless coverage to storage to GPU compute, using token rewards. The model has been debated for years, largely because many networks struggled to show customers paying for the service. A project reporting recurring revenue from AI labs is a notable data point for the sector, similar to how we saw real usage metrics shift the conversation when Solana DEX volume surpassed the NYSE.

Risks to keep in mind

• Self-reported figures. The revenue numbers come from the project and reports citing it, not from audited financial statements.
• Token dynamics. Sharp rallies on funding news can reverse, and any upcoming token unlocks can add supply. Our monthly token unlock roundup is a good habit to check before trading newer tokens.
• Legal and platform questions. Web data collection for AI is an area of active legal and policy debate, and residential proxy networks can face scrutiny from websites and regulators.
• Customer concentration. Revenue tied to a small number of large AI labs could be sensitive to any single customer changing course.

Grass is one of the clearer tests yet of whether DePIN networks can turn token-incentivized infrastructure into a real business. The next few quarters of reported revenue will say more than any funding announcement.

This article is for informational purposes only and does not constitute financial advice.