375ai

What 375ai Is and the Problem It Solves

AI systems that operate in the physical world—autonomous vehicles, smart cities, robotics, retail analytics—require continuous, structured data about how people, vehicles, and commerce actually move through real spaces. The problem is that most AI training data consists of sparse photographs, short video clips, and text scraped from the internet. The physical world, where 75% of commerce happens offline, remains largely invisible to AI models.

375ai addresses this gap by operating as a decentralized physical infrastructure network (DePIN) that continuously captures, processes, and delivers structured real-world observations. It positions itself not as an end-user application but as foundational data infrastructure—"an ingredient, not an application"—supplying the annotated, machine-readable ground truth that downstream AI systems require.

How It Works

The 375ai network has two primary data-collection layers: hardware nodes and a mobile application.

375edge is the company's enterprise-grade edge node. Each unit is equipped with six high-definition cameras, audio sensors, environmental sensors, and an on-device NVIDIA Jetson GPU. Rather than streaming raw footage to the cloud, 375edge performs on-device inference: it ingests multimodal inputs—video, audio, environmental readings—and outputs only structured metadata such as vehicle counts, speeds, traffic incident classifications, pedestrian density, and movement vectors. This edge-first design reduces bandwidth costs, enables real-time latency, and ensures that personally identifiable information (license plates, faces) is stripped before any data leaves the device.

375go is the companion mobile application. Users passively contribute data—network signal strength, location-stamped observations, and environmental readings—while going about their normal routines. The app maps contributions across hexagonal geographic cells, creating crowd-sourced coverage that complements the fixed node network.

Data produced by both layers is processed on-chain through the Solana network, which provides the transaction throughput and low-cost settlement needed for high-frequency, real-time data attestation. Solana's architecture is well suited to the DePIN use case: microsecond finality and sub-cent transaction fees allow the network to record and settle data contributions at a cadence that would be economically unworkable on higher-fee chains.

Token Economy

The network's native token, $EAT (Edge AI Token), has a total supply of 1 billion and serves as the unit of value flowing between data producers and data consumers.

Contributors—whether operating a 375edge device at a fixed location or running the 375go app on mobile—earn $EAT in proportion to the quality, accuracy, and uptime of their data contributions. Staking is built into the system: participants stake tokens to validate network integrity and deter spam or low-quality submissions.

On the demand side, businesses and AI developers purchase structured data using fiat currency, which is converted to data credits. Accessing those credits requires burning $EAT, creating continuous buy-and-burn pressure on the token. This mechanism attempts to tie token value directly to data consumption rather than speculation alone.

The published token allocation is: 40% community and contributor rewards, 25% team, 20% private investors, 4% marketing and partners, 3% public sale, 3% liquidity, 3% treasury, and 2% additional liquidity reserves.

Scale and Deployment

At mainnet launch in October 2025, 375ai reported observing over 2 million vehicles per day through its deployed 375edge devices. The network has a strategic partnership with Outfront Media (NYSE: OUT), giving it access to 40,000 premium high-traffic locations across the United States—covering approximately 70% of the US population. Fixed deployments are live in Los Angeles, with expansions underway in New York and Miami.

The 375go mobile app has surpassed 300,000 downloads across 170+ countries, providing broad geographic coverage beyond the hardware node footprint.

Key Features

• Privacy-by-design: Personal identifiers are anonymized at the point of capture on the edge device. No raw video or identifiable footage is transmitted or stored.
• Multimodal data: The network captures traffic flow, environmental conditions, audio signals, and connectivity data, making it broadly applicable to use cases beyond any single vertical.
• Enterprise hardware: 375edge nodes use NVIDIA Jetson compute, designed for continuous outdoor deployment at commercial scales.
• Real-time output: Edge inference delivers structured insights with minimal latency, enabling time-sensitive applications in autonomous systems and smart infrastructure.
• Open data economy: Structured datasets are available for purchase by AI developers, government agencies, urban planners, logistics operators, and other enterprise consumers.

Funding and Team

375ai raised a total of $10 million across multiple rounds, with the most recent round announced in October 2025. Lead investors include Delphi Ventures, Strobe Ventures, and HackVC, with participation from 6MV, ARCA, EV3 Ventures, Peaq, and Heartcore Capital. The Delphi involvement is notable given Delphi's track record investing in foundational Solana ecosystem projects.

Harry Dewhirst serves as CEO. Chad Partridge holds the role of Chief AI Officer. No security audits of on-chain contracts had been publicly disclosed as of the mainnet launch.

Solana Ecosystem Fit

375ai represents one of the more data-intensive DePIN deployments on Solana. The project's design requirements—high-frequency data writes, micropayment settlement for thousands of simultaneous contributors, and a token economy requiring low-latency on-chain operations—align closely with the properties that make Solana competitive among smart contract platforms. Its Outfront Media partnership gives it immediate real-world coverage that many DePIN projects must build from scratch over years.

The project occupies a distinct niche compared to most Solana DePIN networks, which tend to focus on wireless coverage (Helium), GPU compute (io.net), or mapping (GEODNET). 375ai's focus on structured real-world observational data—specifically the kind of dense, continuous ground-truth data required for training and validating physical AI systems—positions it as complementary infrastructure for the broader wave of AI development.