Grass’s data infrastructure has also expanded significantly. The multimodal repository now exceeds 250 petabytes of content, a volume sufficient to train a frontier-scale video generation or speech model. Supporting this scale required an intense focus on structural efficiency. Grass successfully established what it believes is the lowest sustainable cost basis for storing and retrieving data within the residential proxy sector. This foundation, combined with an ethical approach to network growth, allows Grass to price its products competitively while delivering consistent quality across various workloads.

Progress continues on Grass’s early real-time data products. A production-ready Search Engine Results Page (SERP) API is currently in trials with three major SEO firms and one leading AI lab, receiving positive early feedback. Internally, this API is considered version zero of Live Context Retrieval (LCR), representing the crucial first step in enabling real-time web context for AI models. Grass anticipates LCR will become an increasingly meaningful segment of the business throughout the year ahead.

Mobile adoption played an important role in that expansion. The Grass mobile app, launched earlier this year, has steadily grown its share of total network activity, today, approximately 38% of the network is on mobile. This shift contributes meaningfully to the size and geographic reach of the network.

Business Overview

Grass has grown from a concept into a revenue-generating network with accelerating performance. It is now among the most widely adopted crypto protocols and a core data infrastructure provider for leading AI companies.

At the start of the year Grass began to actively market its products and services to potential customers for the first time. Grass quickly grew from having close to no meaningful revenue in Q1 of 2025 to approximately $2.75m in Q2. In Q3 Grass generated approximately $4.3M in revenue. In Q4 Grass is projected to generate approximately $12.8M in revenue. These projections are based on an expected revenue of $10M in October and November 2025 based on ongoing business. Certain revenues generated in Q4 are also the result of contracts signed in Q3 with enterprise customers. Grass has been and will continue to reinvest the majority the revenue it generates into growth.

Several commercial milestones shaped 2025 revenue. Earlier in the year, Grass closed enterprise contracts with two major AI organizations, which helped establish the foundation for subsequent growth. In Q3, a channel partner began white labeling Grass’s video downloader. In Q4, Grass closed additional enterprise agreements with a large AI organization and a leading lab in the video generation space. Grass has also seen repeat business from nearly every AI customer, reflecting consistent performance and reliability.

Grass does not disclose the names of AI labs purchasing pre-training data. Training datasets are considered proprietary intellectual property across the industry, and disclosure practices are being shaped primarily by the labs in consultation with regulators. Grass’s role is to operate reliable, confidential infrastructure that supports those standards.