Understand RoboFi in one article and understand the new star projects on the Web3 robot ecosystem track

Table of contents

• 1.@openmind_agi

• 2.@peaq

• 3.@GEODNET_

• 4.@psdnai

• 5.@PrismaXai

• 6.@NRNAgents

• 7.@AukiNetwork

• 8.@RoboStack_io

• 9.@frodobots

• 9.1. @BitRobotNetwork

• 9.2. @ET_Fugi

• 9.3. @samismoving

• 9.4. @robotsdotfun

• 9.5. @UFBots

• Summarize

1.@openmind_agi

Keywords: operating system, decentralized collaboration layer

OpenMind received a $20 million financing and is committed to enhancing robots' cognitive and learning capabilities through open source architectures. Its core products cover:

OM1: An AI-native operating system, known as "Android in the field of robots", is hardware-independent and completely open source. It gives robots the ability to perceive the environment, store memory, formulate strategies and execute actions, allowing them to understand the world around them and respond in a contextual manner. Previously, some developers used OM1 in Hackathon to create an intelligent robot for care for Alzheimer's patients within 48 hours.

FABRIC: As a decentralized collaborative network layer, blockchain technology is used to realize data sharing and rapid learning between robots (such as language acquisition), supports consensus mechanisms, task execution and result settlement, and promotes multi-machine collaboration intelligence.

2.@peaq

Keywords: Layer 1. Machine economy ecology

Peaq completed a US$21 million financing and is a Layer 1 public chain designed specifically for DePIN and machine economy.

This project allows robots, drones, sensors and other physical devices to have self-savereign IDs on the chain, realize point-to-point transactions, automated operations and data storage, support scenarios such as robot leasing, unmanned delivery, and multi-equipment collaborative operations, and build a truly "machine-active economy" - allowing devices to independently participate in economic activities like AI agents.

Based on Substrate development, compatible with the Polkadot ecosystem, it has been connected to more than 90 blockchains. There are more than 50 DePIN projects active on the peaq chain, such as Silencio (noise monitoring), MapMetrics (driving as earning), Farmsent (agricultural traceability) and Natix (city perception). The entire network is connected to more than 4.5 million physical equipment, serving more than 20 industries including energy, transportation, agriculture, and environment.

3.@GEODNET_

Keywords: High-precision positioning network

GEODNET received a $15 million investment, focusing on building the world's largest decentralized real-time dynamic (RTK) positioning network and providing centimeter-level accurate GNSS navigation services.

Originally deployed in Polygon, it has now expanded to the Solana ecosystem.

Anyone can purchase and install GEODNET's "Satellite Miner" hardware equipment, collect RTK correction data and upload it to the network, and receive a $GEOD token reward. These high-precision positioning data will be widely used in autonomous vehicles, aircraft, robots and even meta-cosmic space positioning.

At present, the network has covered 145 countries, established more than 19,000 sites, and has transmitted over 207,178 GB of RTK data.

4.@psdnai

Keywords: Real-world data collection

Poseidon was hatched by @StoryProtocol, and the seed round raised $15 million to solve the problem of insufficient physical data required for robots and multimodal AI Agents to train in real environments.

The project builds a decentralized data infrastructure, relying on Story Protocol to collect, organize and manage multi-source data from the real world (including video, audio, images, geographic location, etc.), and ultimately outputs a high-quality training set protected by intellectual property rights to accelerate the iteration of AI models.

The APP is currently online, with voice data as the core in the early stage, and users are encouraged to upload voice clips under different accents, dialects and background noise, helping AI improve voice recognition capabilities in complex environments.

5.@PrismaXai

Keywords: Tele-op remotely control, training data generation

PrismaX was led by a16z CSX, with a total financing of US$11 million, with the goal of accelerating the robot's move toward a high degree of autonomy through human intervention.

Its tele-op platform allows users to directly control the robotic arm through the browser, and all operational processes are recorded as high-fidelity behavior data for training and optimization of AI models. Operators can obtain platform token incentives.

In the future, we plan to connect to mainstream robot brands such as Yushu and UBL.

The short-term focus is on remote operation and visual data acquisition;
Complete actual tasks through remote command of manpower in the medium term;
The long-term vision is to achieve fully autonomous operation of robots.

6.@NRNAgents

Keywords: remote control, reinforcement learning, AI Agent training platform

NRN Agents is a decentralized AI Agent development and training platform running on Arbitrum. Originally developed from the research and development of the game AI Agent, it has gradually extended to embodied intelligence - with the help of the "Sim-to-Real" methodology, it provides an efficient training path for the AI ​​Agent behind the robot.

Users can control virtual robots through browsers, and the behavioral trajectory generated by each interaction becomes a valuable training material for imitation learning; at the same time, the AI ​​Agent in the system can continue to evolve through reinforcement learning without the need for reinitiative training.

At this stage, we focus on the operation training of RME-1 robotic arm, and will be expanded to the field of robotic competition, such as simulation and actual combat of sports skills such as fighting and track and field.

7.@AukiNetwork

Keywords: DePIN, spatial perception, physical data sharing

Auki Network is committed to creating a decentralized "collective nervous system" for robots and smart devices to enhance their understanding of the real world. Its core is a spatial data network based on the DePIN mode, which supports various devices (such as robots and AR glasses) to upload location and perceive information in real time, and jointly build a collaborative spatial cognition system.

The network is based on the Posemesh protocol and contains four types of nodes:

• Computing node : Provide computing power resources
• Motion node : From robots or other sensor devices, contributing mobile and environmental data
• Reconstruct nodes : Integrate the first two data to generate 3D maps or spatial models
• Domain name node : responsible for managing and naming 3D spaces

Each node receives $AUKI token rewards based on the degree of contribution.

8.@RoboStack_io

Keywords: software and hardware decoupling, standardized communication protocol

RoboStack is a cloud-native robot simulation and integration platform. It not only provides high-fidelity environmental modeling and elastic computing resources, but also launched the RCP (Robot Context Protocol) protocol to break through barriers between different hardware platforms, software frameworks and communication standards, significantly reducing the complexity of AI Agents connecting with the underlying robot hardware.

The native token $ROBOT has been logged into the virtuals platform. Users can participate in governance by pledging $ROBOT and vote for the robot projects you support in the SIM lord module to win additional rewards.

9.@frodobots

Keywords: sidewalk robot, gamified data collection

FrodoBots Lab is a robotics laboratory focusing on the development of embodied intelligence, promoting real-world data accumulation through gamification. Its main product, Earth Rover, is a remotely controlled sidewalk robot priced at US$249-399, suitable for entertainment, scientific research and AI training. As of now, FrodoBots has deployed hundreds of Earth Rover around the world and has publicly released more than 2,000 hours of real driving data sets.

The laboratory combines innovative hardware, decentralized platforms and AI architecture to create an ecosystem that integrates fun and practicality. The specific projects are as follows:

9.1. @BitRobotNetwork

Keywords: subnet architecture, decentralized platform

BitRobot Network is a Solana-based decentralized network developed by FrodoBots Lab in collaboration with Protocol Labs, and received $6 million in financing. Modular subnet design is adopted, each subnet focuses on specific resource supply, such as computing power, robot fleets, data sets (real or synthetic) or AI models, adapting to diverse forms from sidewalk robots to humanoid robots. The first subnet ET Fuji has been officially launched.

9.2. @ET_Fugi

Keywords: Real game, remote control

ET Fugi is a robot game that combines reality and NFT. Players "catch" alien NFTs in real streets by remotely controlling Earth Rover. The walking trajectory and environmental interaction data generated during the process can be used to train robot AI, and players can win rewards such as FrodoBots Points (FBP).

9.3. @samismoving

Keywords: AI Agent framework, remote control

SAM was originally launched in collaboration with FrodoBots and @virtuals_io to remotely control Earth Rover. It was then developed into the core framework for the release of AI Agent on the Robots.Fun platform. All new Agents are built on the SAM architecture and need to pay a 5% profit share to it.

SAM uses Earth Rover to collect runtime data and feed back to the overall AI capabilities upgrade. In addition, some adventures will be broadcast live through X (Twitter), showing interactive challenges, competitions or data collection processes, both educational and viewing.

9.4. @robotsdotfun

Keywords: AI Agent distribution platform

Robots.Fun is a launch platform specially designed for robot AI Agent, which drives data accumulation through daily mission competitions (such as capturing alien NFTs) to improve the level of AI intelligence. To create an AI Agent on the platform, users must hold or purchase an Earth Rover as their "body".

9.5. @UFBots

Keywords: robot battle, remote control

Ultimate Fighting Bots is a free and open robot fighting community, where users can remotely control humanoid robots based on Booster T1 for real-time duels. The platform generates a large amount of real adversarial data to train embodied AI, and also provides users with an immersive competitive experience.

Summarize

After browsing the above projects, it is not difficult to find that the integration of Web3 and robots shows two core values:

1. With the help of the decentralized characteristics of blockchain, break the monopoly pattern of AI models and training data and lower the entry threshold;
2. The combination of on-chain rights confirmation and token incentive mechanisms stimulates users' enthusiasm for participation and injects sustainable growth momentum into the robot economy.

The above is the detailed content of Understand RoboFi in one article and understand the new star projects on the Web3 robot ecosystem track. For more information, please follow other related articles on the PHP Chinese website!