In his speech at Davos earlier this year, Musk reiterated that provocative prophecy – in the future, the number of robots on Earth will surpass that of humans.

Clearly, AI and robots have essentially become the two main technology topics worldwide: one is the general artificial intelligence that is approaching the AGI critical point, and the other is robots that are emerging from the lab, attempting to take over physical labor from humans. Similarly, aside from the AI concept, the cryptocurrency industry’s key focus this year also includes embodied intelligence. Here are some noteworthy projects in the Robotic track.

OpenMind

On August 4, 2025, according to official news, OpenMind, a smart machine infrastructure company headquartered in Silicon Valley, announced the completion of a $20 million financing round led by Pantera Capital, with participation from Ribbit, Sequoia China, Coinbase Ventures, DCG, Lightspeed Faction, Anagram, Pi Network Ventures, Topology, Primitive Ventures, and Amber Group, among others, as well as several well-known angel investors.

OpenMind helps robots think, learn, and work by developing open-source software. Its native open-source AI robot operating system OM1 allows for the configuration and deployment of AI Agents in both the digital and physical worlds. Users can create an AI character that runs in the cloud, or operate it on physical robots in the real world.

In simple terms, OpenMind is creating OM1, which is like the "AI brain" for robots. This "AI brain" can be co-operated by multiple AI Agents, interact with several LLMs, and source data from various origins for tasks (such as helping users post on social media). Since OM1 is open source, it is a highly adaptable robot operating system, just like how the Android system is hardware-agnostic for smartphones.

Additionally, OpenMind has an on-chain robot identity network called FABRIC, designed to provide a verifiable trust layer shared by humans and robots. Humans can earn badges by sharing location data via maps, assessing robot behaviors, and developing features, while every robot loaded with the OM1 system will join the FABRIC network, obtaining a unique verifiable identity, allowing commands, operation logs, ownership, and relevant actions of robots to be traced on-chain.

In December 2025, OpenMind and stablecoin issuer Circle jointly announced the launch of a robot autonomous payment system based on the x402 protocol. As robots become more capable, they will no longer merely be tools for executing tasks but will start to play the role of autonomous economic entities. They will need to purchase computing power, data, skills, and even hire other robots or humans to complete complex tasks.

CodecFlow

CodecFlow provides a unified platform that seamlessly operates across cloud, edge, desktop, and robot hardware while supporting currently popular APIs and traditional systems. The platform normalizes the input from different robot sensors into a common format and modularizes relatively complex robot actions, allowing development teams or users to design robots without starting from scratch, enabling perception, decision-making, and control among robots to mutually influence through the network, rather than fragmenting into a single platform specific to hardware.

AI-driven operators respond to UI changes in the software or variations in the robot's environment by perceiving and reasoning in real-time, addressing the vulnerability of traditional robotic automation processes, which overly rely on pre-written scripts when faced with even slight changes. In short, it captures screenshots, camera footage, or sensor data, then uses AI to process these external input data to handle observations or instructions, ultimately executing decisions through user interface interactions.

Peaq

On March 27, 2025, the DePIN Layer1 protocol Peaq completed a $15 million financing round led by Generative Ventures and Borderless Capital, with participation from Spartan Group, HV Capital, CMCC Global, Animoca Brands, Moonrock Capital, Fundamental Labs, TRGC, DWF Labs, Crit Ventures, Cogitent Ventures, NGC Ventures, Agnostic Fund, and Altana Wealth.

Although it initially focused on the DePIN narrative, Peaq launched the Robotics SDK in September last year, enabling robots to acquire autonomous identity, conduct payments and receipts, validate data, and integrate into the on-chain network economy. Now, any robot that is compatible with the ROS2 system can join the Peaq network economy, using its universal standards to transact with humans or other robots.

Additionally, last year, Peaq launched a robot RWA project called "RoboFarm" on DualMint, where they established a robotic farm in Hong Kong, achieving 80% automation in agricultural production through robots. The resulting lettuce, spinach, and kale are sold in Hong Kong. The expected annualized return for NFT holders is about 18%.

Axis Robotics

Axis Robotics is dedicated to building a distributed scalable infrastructure for embodied intelligence (Physical AI). They firmly believe that a simulation-first approach is the best path to overcome the bottleneck of robotic data scarcity and model generalization, achieving a triple leap in the quality, richness, and scale of data through low-cost, large-scale data collection combined with a unique data enhancement engine. Moreover, each data asset has a credible on-chain provenance, collectively constructing the core fuel repository that drives the evolution of general robot intelligence (RGI).

Axis has innovated the way robotic training data is provided. Most other projects that "input/offer robot training data" mobilize users to capture and upload videos of specified actions completed in reality using devices like smartphones and smart glasses, achieving low-barrier, global participation. Although this method incurs low data collection costs, the physical realism of video-captured data is insufficient, lacking depth information and cannot guarantee the continuity and accuracy of 3D data.

Through "simulation," Axis addresses this pain point by employing a vast array of diverse simulated environments (lighting, angles, friction, dynamics, etc.) to allow models to accomplish tasks in more demanding virtual conditions, thus acquiring strong generalization capabilities. Axis adopts a Hybrid Strategy, combining scarce real data with vast amounts of synthetic data. Utilizing GPU-accelerated metadata enhancement technology, it achieves a large number of variations in lighting, texture, and physical attributes for a single scene. The virtual scene is not static or hard-coded; it can be flexibly adjusted. Countless scenarios can be generated with code, allowing robots to confront more rigorous and comprehensive challenges under various scene requirements. The cost of generating scenes is low, while the output quantity is enormous. This effective method of converging using large amounts of data to approach optimal solutions has been partially validated by industry giants like Google and NVIDIA.

Axis has publicly opened its first simulation robot learning project, "Little Prince's Rose." In this project, users can remotely control robots through the web to successfully water plants in a simulated environment, collecting and analyzing user operations to teach the robot how to water plants. Users can operate the robot via the webpage, maintaining the low-cost, low-barrier nature of video upload collection, while building a native 3D-aware VLA (Vision-Language-Action) foundational model for the robot, enhancing its spatial thinking ability that is lacking from video data input channels.

Just five days after its launch, ordinary users worldwide without a background in the robot industry contributed tens of thousands of high-quality, effective trajectories for strategic training through an engaging experience in the "Little Prince's Rose" project. Based on this data, Axis successfully trained a strategy model and reproduced the real machine of the Franka robotic arm. This marks that Axis has successfully run through the full-stack closed loop of "task generation -> community collection -> data enhancement -> model training -> real machine deployment."

One hour of real data can be converted into 1,000 hours of training data, significantly reducing the cost required for the generalization of robotic models.

During the Spring Festival's Beta testing, again within just five days, 18,000 participants without a robot industry background completed 27 brand new tasks on Axis, contributing over 100,000 data trajectories. The successful test supported high levels of intra-task randomization and verified compatibility with diverse assets such as wheeled robots and dual-arm robots.

Axis's core product will officially launch in late March and plans to open-source the world's largest purely simulated dataset based on the Franka robotic arm by the end of April or early May, fully meeting the needs of strategy and model training. Simultaneously, as a robotics track project originating from Crypto-AI, Axis has begun exploring and promoting external industry landings, accelerating commercialization processes around several benchmark clients in various segmented fields: collaborating with an automotive company to promote the landing of automated production solutions; reaching a consensus on cooperation in virtual assets and world modeling with a quasi-IPO computing power company; and establishing deep cooperative relationships with many companies in the embodied entity sector for virtual simulation data collection and model training in critical areas. These reflect the rare externality of Crypto projects.

GEODNET

GEODNET is a decentralized network providing real-time dynamic positioning data with centimeter-level accuracy for drones, robots, and others, boasting over 21,000 active stations across more than 150 countries. Over the past year, the project has generated revenue exceeding $7 million, with a trend of quarterly growth.

Although the project is often categorized under DePIN, the growing application of robotic technology in real life is expected to create a broader demand for high-precision real-time positioning data. In February 2025, Multicoin announced that it would lead the acquisition of $8 million worth of $GEDO tokens from the GEODNET Foundation.

BitRobot

The BitRobot Network is developed collaboratively by FrodoBots Lab and Protocol Labs, aimed at achieving distributed robotic work and collaboration. Its key components include: verifiable robot work (VRW, a quantified metric for network rewards) for defining and validating robot tasks, equipment node tokens (ENT, the unique identifier for robots within the system, existing in NFT form) for device ownership and network access, and a subnet that serves as the execution layer for task operations (resource clusters that create value for the BitRobot network).

On February 14, 2025, FrodoBots Lab announced the completion of a $6 million seed round financing, bringing the total financing amount to $8 million.

FrodoBots Lab also sells robots; the Earth Rovers are like real-life Mario Karts, priced at $249, allowing players to remotely control their robots in a global treasure hunt game called ET Fugi through a browser, with data being provided for researchers to deploy and test their latest AI navigation models. ET Fugi is also the first subnet of BitRobot.

Another gaming robot, Octo Arms, will also be launched in the future, allowing players to remotely control robotic arms to complete various 3D puzzle games and competitions.

The concept of a "subnet" in this robotic network is somewhat abstract; in simple terms, any cluster that can contribute to the overall network ecosystem (or specific projects/events that the cluster undertakes) is a subnet, such as the ET Fugi game mentioned above and SeeSaw launched by Virtuals.

SeeSaw

SeeSaw is the 5th subnet of BitRobot, a robot training data sharing application launched by Virtuals last October. In SeeSaw, users take videos of their daily behaviors, upload them to complete tasks and earn rewards. These video data from global users, including activities like tying shoelaces and folding clothes, will be used to train robots.

Auki

Auki’s decentralized machine perception network Posemesh connects humans, devices, and AI. At its core is a DePIN (decentralized entity network) architecture, allowing robots, AR glasses, and other devices to share location and sensor data in real-time, collaboratively building a spatial understanding of the physical world, and providing a shared spatial view for robots, AR, and AI.

Various node roles are designed based on the Posemesh protocol. Compute nodes provide computational power, movement nodes (robot terminals) upload location information and sensor data, reconstruction nodes generate 3D map models based on this data, and domain nodes manage the 3D space. Each node is rewarded with $AUKI tokens based on its contribution, driving a self-evolving machine vision network.

This network emphasizes privacy protection, preventing any single entity from monitoring users' private spaces, while being applicable in various scenarios such as retail (for product placement optimization), property management (for asset tracking), and exhibition navigation and construction renovation.

Their Cactus AI spatial computing platform has already initiated positive pilots with Toyota Material Handling and Swedish supermarket Stora Coop.

XMAQUINA

A DAO that enables retail investors to participate in robot enterprise investments. The DAO raised $10 million by selling its $DEUS tokens in batches. Currently, this DAO has used the auction proceeds to acquire shares in six robotics companies, including Apptronik, Figure AI, Agility Robotics, 1X Tech, NEURA Robotics, and Robotico, with some investments already yielding returns, even surpassing 100% in some cases.

PrismaX

On June 17, 2025, PrismaX announced the completion of a $11 million financing round, with investors including a16z CSX, Volt Capital, Blockchain Builders Fund, Stanford Blockchain Accelerator, and Virtuals.

PrismaX is building an open coordination layer connecting remote operators, robot users, and robot companies. Operators can connect with users to remotely control robots and complete actual tasks while gathering valuable data. They can also request real-world services such as logistics and advertising.

PrismaX also has a protocol for remote-operated robots, where companies can seek experienced robotic operators capable of handling complex tasks. Operators can choose to stake network tokens to enhance their credibility and increase opportunities for high-yield tasks. Rewards for stakers are not only related to their staked amount but also depend on their work quality and can earn extra rewards as their work efficiency improves.

The data accumulated from remote operations will be used to train robots to enhance their autonomy, which will in turn improve the work efficiency of remote operators, ultimately achieving a high level or even complete autonomy for robots.

NRN Agents

NRN evolved from the AI Agent battle real-time training chain game AI Arena. On October 28, 2021, developer ArenaX Labs announced it had completed a $5 million seed round financing led by Paradigm Capital, with Framework Venture Partners participating. On January 9, 2024, ArenaX Labs announced the completion of a new funding round of $6 million, led by Framework Ventures, with participation from SevenX Ventures, FunPlus/Xterio, and Moore Strategic Ventures.

Although the general process is still data collection -> reinforcement robot learning, NRN leverages its rich experience in the gaming field to provide a browser-based experience converting robotic data collection into a game, allowing users to intuitively control simulated robots through a browser. During the gameplay, the user-generated behavior data is used to train real-world robotic systems.

In the current phase, the project will focus on robotic arms (RME-1) to validate data collection, real-time learning, and adaptability.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。