Abstract

In 2026, the deep integration of AI, robotics, and Web3 infrastructure is giving rise to a brand new narrative - the Web3 Robots track. This track has moved beyond mere conceptual hype and has entered a crucial phase of real-world application. This research report systematically outlines the definitions and developmental logic of the Web3 Robots track, analyzing representative projects such as OpenMind, PrismaX, peaq, Virtuals Protocol, Geodnet, and XMAQUINA from three dimensions: infrastructure layer (operating system, data layer, network layer), machine economy layer (tokenized platforms, positioning networks), and application implementation layer (DePIN investment). The research finds that this track is forming a three-layer collaborative architecture: "data layer trains intelligence - system layer unifies standards - network layer incentivizes collaboration," where machines are transitioning from closed tools to autonomous economic entities with on-chain identities, payment capabilities, and collaborative networks. Although it is still in the early stages focused on building underlying capabilities, the emergence of real income cases (such as the peaq ecological robot farm with 18% APY) indicates that the track has progressed from concept to a tangible, observable, and participatory development stage. For investors, focusing on the actual deployment scale of projects, their ability to close the commercial loop, and the quality of institutional endorsements will be key to capturing the next hundredfold opportunity.

1. Track Definition and Evolution Logic: From Automation Tools to On-Chain Economic Entities

Robot technology is not a new phenomenon. Over the past decade, industrial robotic arms, warehouse robots, drones, and other automation devices have been scaled in manufacturing and logistics. However, these robots are essentially still tools within closed systems — they execute preset instructions but lack capabilities for identity recognition, autonomous decision-making, value exchange, and cross-platform collaboration. With the AI large models bestowing "thinking" abilities upon machines, and blockchain technology providing identity and settlement infrastructure, a new paradigm is forming: machines are no longer mere hardware but can possess on-chain identities, autonomously complete transactions, and participate in real-world production as autonomous economic entities.

The core driving force behind this transformation comes from technological maturity across three levels. First, the breakthrough in embodied intelligence (Embodied AI). Large language models and multimodal models enable robots to comprehend natural language, perceive their environment, and plan tasks. For example, OpenMind's OM1 operating system integrates perception, memory, reasoning, and action into a unified framework, allowing robots to evolve from "being able to move" to "being able to understand and think." Second, the rise of decentralized physical infrastructure networks (DePIN). Blockchain provides identity authentication (DID), trustworthy data recording, and automated settlement capabilities for physical devices, allowing machines to participate in market transactions as economic entities. Third, the maturation of stablecoins and Layer 2. Efficient micro-payment infrastructure makes high-frequency and small-value settlements between machines feasible, laying the financial foundation for the machine economy.

Forbes' 2026 forecast noted that blockchain would become the trust network for AI, where every significant agency action would be written into a lightweight ledger, achieving compliance, governance, and accountability. This implies that the Web3 Robots track is essentially constructing a new economic system that allows machines to prove themselves, trust others, create value, and participate in distribution. This system consists of a three-layer architecture: at the bottom is the operating system that endows machines with intelligence (like OpenMind OM1), the middle is the protocol layer that provides identity and collaboration networks (like OpenMind FABRIC and peaq), and at the top is the labor market and tokenized platforms focused on application scenarios (like Konnex and Virtuals). The synergy of the three-layer architecture is reshaping robots from "tools" to "digital citizens."

2. Infrastructure Layer: Synergistic Evolution of Operating Systems, Data, and Networks

Construction of the underlying infrastructure for the Web3 Robots track is advancing in multiple dimensions, with the most representative being OpenMind's layout in the operating system layer, PrismaX's exploration in the data layer, and peaq's construction in the network layer. These three projects form a complete infrastructure closed-loop of "system - data - network," providing the operational basis for upper-layer applications.

OpenMind is hailed as the "Android in the robotics field." Its core products include the open-source AI-native robot operating system OM1 and the decentralized collaboration network FABRIC. OM1 is licensed under MIT, has received over 2,500 stars on GitHub, attracting more than 500 global contributors and over 7,500 independent developers. Unlike traditional robotic operating systems (ROS), which only focus on motion control and navigation, OM1 integrates four major modules: perception, memory, reasoning, and action, supporting advanced functionalities like natural language interaction, environment mapping, and object recognition. Currently, OM1 has been adapted by over 10 leading hardware manufacturers including Unitree, Fourier, UBTECH, and Deep Robotics, covering various forms such as humanoid robots, quadrupeds, and robotic arms. The FABRIC protocol constructs a decentralized machine collaboration network, assigning on-chain identities (peaq ID) to each robot, enabling skill sharing, task coordination, and USDC micro-payment settlements among machines. In February 2026, FABRIC Protocol (ROBO) launched on Binance Alpha and Binance Futures, with a 24-hour trading volume exceeding $140 million, subsequently listed on major exchanges like OKX, Coinbase, and Kraken. The project completed about $20 million in financing in August 2025, led by Pantera Capital, with participation from top institutions like Coinbase Ventures, DCG, and Sequoia China, and the latest round valuation is about $200 million, while the Kaito Launchpad presale valuation reached $400 million FDV. Current participation points include Season 1 points program, NFT minting for the FABRIC Identity Network, and GitHub code contributions, with strong expectations for airdrops.

PrismaX is hailed as the "gold mine" of training data for the physical world. If algorithms are the "brain" of robots, data is the nourishment for the "brain." PrismaX is positioned at the AI robot data layer, using a human-machine collaboration (RLHF) model to solve the scarcity of "physical world interaction data" in robot training. Its platform allows users to remotely control real robotic arms to complete actions via a web interface, recording operational data for sale to robot companies for AI training, while users earn points to redeem future tokens. This “Play-to-Train” model creates a data flywheel: more users participating generates more data, leading to better training models, which in turn attracts more users. PrismaX recently completed a $11 million seed round led by top venture capital a16z, with participation from Virtuals Protocol. Currently, the ecosystem has over 500 participants completing remote robotic arm operations, having launched two complete operable robotic arm systems (Unitech Walker "Tommy" and "Bill"). Users can earn points through daily check-ins, white paper quizzes, or paid training ($99), with market expectations for future airdrop distributions. The risk is that a large influx of “farming studios” may dilute point value, and whether remote operation data can truly train commercial-grade robots remains a matter of industry debate.

peaq is the Layer-1 network for the machine economy. peaq is a Layer-1 blockchain specifically designed for the machine economy, with core functions including machine identities (peaq IDs), on-chain wallets, access control, and nanosecond-level time synchronization, supporting millions of robots and devices for autonomous transactions. Unlike many DePIN projects still in the conceptual stage, peaq has successfully verified a real business closed-loop. Its ecological Hong Kong robot farm (Robo-farm) uses automated robots to grow hydroponic vegetables, where users purchase NFTs representing a share of the farm, and income from vegetable sales is converted to stablecoins and distributed directly on-chain to NFT holders. The first income distribution at the end of January 2026 showed a single payment of 3,820 USDT, with an annualized return of approximately 18%. This "profit from selling vegetables without relying on token inflation" model has made peaq a typical case for the realization of RWA (real-world assets). As for partnerships, peaq has engaged in technical validation collaborations with industrial giants like Bosch, Mastercard, and Airbus, covering fields such as IoT sensor integration, payment gateway integration, and supply chain tracking. The mainnet went live in 2024, currently with a circulating market cap of about $34.25 million and FDV around $78 million, with 50-60 DePIN applications operating within the ecosystem, connecting over 2 million physical devices. The token $PEAQ is mainly used for Gas and staking, with the "Get Real" campaign ongoing, offering a reward pool of 210 million $PEAQ (worth over $10 million), where users can earn XP/NP through completing real-world DePIN tasks and receive tokens.

The relationship among these three is akin to a complete production system: PrismaX provides the data "raw materials" to train robots, OpenMind's OM1 provides the "operating system" for robots to run intelligently, and peaq provides the "network and incentive layer" for robots to complete economic settlements. Their collaboration forms a complete infrastructure stack for decentralized embodied intelligence.

3. Financial Layer of the Machine Economy: Tokenized Platforms and Positioning Networks

Once the infrastructure layer addressed the issues of "how machines become intelligent" and "how machines collaborate," the financial layer of the machine economy began to emerge. The core problem addressed at this layer is: how to price, trade, and circulate the value of machines? Virtuals Protocol and Geodnet provide answers from different perspectives.

Virtuals Protocol is a tokenized platform for AI agents/robots, allowing community participation in the issuance, staking, and governance of agents. Its core mechanisms include the Pegasus/Unicorn ecosystem, the ACP (Agent Commerce Protocol) market, and Butler tools. The ACP market allows for trustless business transactions between AI agents, supporting end-to-end on-chain task publishing, verification, and settlement. In March 2026, Virtuals collaborated with the Ethereum Foundation dAI team to develop the ERC-8183 standard (Agentic Commerce), introducing work primitives with on-chain custody, evaluator certification, and modular hooks to enable trustless business transactions among agents. Data show that on-chain income among agents on Virtuals has surpassed $3 million (excluding transaction fees), marking the formation of a verifiable economic output entirely generated by AI agents. The token $VIRTUAL was launched at the end of 2023, currently valued at around $500 million, and has been listed on major CEXs like Gate.io. The weekly Epoch airdrop system remains active, distributing rewards based on veVIRTUAL staking and Butler usage, with 2% for stakers and 3% for ecosystem participants. At the beginning of 2026, Virtuals also reached a collaboration with OpenMind's FABRIC protocol, allowing robots to obtain economic identities through FABRIC, receive tasks from agents via ACP, and conduct on-chain settlements, realizing a deep integration of the machine economy layer with the infrastructure layer.

Geodnet is referred to as the centimeter-level navigation infrastructure for robots. Geodnet is a decentralized high-precision positioning network built on Solana, providing RTK (Real-Time Kinematic) centimeter-level navigation services for robots, drones, and autonomous vehicles. Its network consists of globally distributed reference stations where node operators earn $GEOD tokens by deploying hardware, while users subscribe to access positioning data. Geodnet's business model features typical "real income" characteristics: 80% of data revenue is used for the buyback and destruction of $GEOD, forming a deflationary mechanism. At the CES exhibition in January 2026, Geodnet showcased the Geoswarm home security drone that can automatically take off from a compact dock on a household rooftop, utilizing GEODNET's high-precision positioning data to complete patrols and return to land automatically without human intervention. In addition, Geodnet also launched consumer-oriented RTK hardware ($150) and RTK measurement receivers ($695), with the latter winning the CES Innovation Award. Geodnet has achieved a total of over $15 million in financing, including rounds led by Multicoin Capital. The token has migrated from Polygon to Solana and is currently tradable on Coinbase. For investors, Geodnet's buyback and destruction mechanism and real hardware sales form a value support, worthy of attention as its node deployment revenue and staking reward mechanisms remain actively operational.

From the perspective of the financial layer, Virtuals Protocol resolved the "liquidity" issue for AI agents — allowing the capabilities of agents to be tokenized, traded, and priced; Geodnet addressed the "spatial awareness" issue for robots — enabling machines to precisely locate and navigate in the physical world. Together, they expand the boundaries of the machine economy: the former allows the value of machines to circulate in the digital world, while the latter enables machines to operate more precisely and reliably in the physical world.

4. Application Implementation Layer: From DePIN Investment to Real-World Assets

The infrastructure layer and financial layer constitute the "skeleton" and "blood" of the Web3 Robots track, but what truly determines the track's vitality is whether the application implementation layer can create value in the real world. XMAQUINA and Robonomics explore this proposition from different dimensions.

XMAQUINA is a robot investment bank governed by DAO. XMAQUINA is a DePIN project that invests in and tokenizes real humanoid robot companies through DAO governance, allowing token holders to share in the profits of robot companies. Its core mechanisms include a "Robotics Bank" and a Machine Economy Launchpad. The DAO allocates capital to promising robot companies (such as Apptronik and Figure AI) and utilizes a SubDAO mechanism for specialized management. In January 2026, XMAQUINA completed its final public auction, with the community contributing over $3.25 million in under 30 minutes, totaling $10 million in financing. The TGE of the token $DEUS is expected to activate between January and February 2026, with 33% unlocked at TGE and 67% released linearly. Current participation points include holding $DEUS for governance voting and profit sharing, focusing on DAO proposals and staking mechanisms, as well as upcoming Launchpad projects. XMAQUINA's model is essentially a "robot version of an investment fund," reducing the participation threshold for ordinary investors in early investments in robot companies while achieving decentralized investment decisions through DAO governance.

Robonomics is the earliest Web3 robot coordination platform. Robonomics is a pioneer in the Web3 robotics space, launching its testnet in 2018, providing robot cloud services and smart contract task allocation functionalities. Its core capabilities include IoT device integration, sensor data on-chain, and automated task execution. The token $XRT of Robonomics was issued in 2019 and is listed on exchanges like Kraken, but its market cap is relatively small, making it an "old veteran" project within the track. Compared to new projects, Robonomics's ecosystem is more mature but lacks momentum, with no large-scale recent airdrops or incentive activities, making it more suitable for long-term investors focused on IoT and robotics integration to hold and observe.

Notably, the application implementation layer is seeing more innovative models. The "Universal Basic Ownership Pilot" project within the peaq ecosystem explores the inclusive ownership of machine assets, while the tokenized machine deployment mechanism allows ordinary users to invest and share in the operational profits of robots. Additionally, the Virtuals ecosystem has seen the emergence of agents like ArAIstotle ($FACY), achieving 382,000 queries, 8,000 users, and $760,000 in tax revenue, with monthly growth in ACP reaching 413 times, proving the immense potential of the AI agent economy.

5. Challenges, Risks, and Future Prospects

Although the Web3 Robots track shows immense imaginative space, it is still in the early stages of development, facing multiple challenges and risks.

On the technical front, the reliability of robot hardware and adaptability to environments remain bottlenecks. As OpenMind founder Jan Liphardt noted, the reliability of critical components, such as dexterous hands, is still a challenge; a mechanical hand with five fingers and 12 degrees of freedom may lose its practical value if it malfunctions after a hundred hours of operation. Furthermore, the gap between simulation tools and real environments, and the need for realistic human voice interaction simulations in social robots also require ongoing technical breakthroughs.

In terms of valuation, some projects face the risk of high valuations with low circulation. Taking OpenMind as an example, the Kaito Launchpad presale valuation reached an FDV of $400 million, doubling from the previous round's valuation ($200 million), which might overdraw secondary market space and face sell pressure from early VC unlocks. Investors should be cautious of projects with "narrative premiums" that are too high while the actual implementation progress falls short of expectations.

On the data quality front, data layer projects like PrismaX face the risk of an influx of "farming studios." If project teams cannot effectively filter high-quality training data, the points may lose value, ultimately leading to severe sell pressure during airdrops. How to incentivize user participation while ensuring data quality is a challenge all data layer projects need to address.

In terms of competitive landscape, traditional robot manufacturers tend to adopt closed systems (like Tesla Optimus), similar to the model of Apple iOS. Whether an open-source "Android model" like OpenMind can survive in the crevices of giants depends on the ability to attract enough mid-tier hardware manufacturers to form a collective ecosystem.

Looking ahead, the development of the Web3 Robots track will evolve along three main lines: first, standardization. A2A (Agent-to-Agent) communication protocols are becoming the universal language for robots and agents, similar to how HTTP unified the early internet, A2A will become the communication foundation of the autonomous world. Second, real income. The case of the peaq robot farm demonstrates that Web3 robot projects can produce real cash flows that do not rely on token inflation. More projects will explore business models like "Device-as-a-Service" and "Robot-as-an-Asset" in the future. Third, compliance and governance. As robots become deeply involved in economic activities, regulatory bodies will demand that AI decision-making be explainable and traceable. The immutable ledger of blockchain will become a key infrastructure to meet compliance requirements.

6. Conclusion: Participation Strategy and Investment Logic

The Web3 Robots track is at a critical stage of transition from proof of concept to large-scale application. For investors and ecosystem participants, the current core strategy should be: focus on building underlying capabilities, track actual deployment scales, and seize early participation opportunities.

From an ecological perspective, infrastructure layer projects (OpenMind, peaq) have higher certainty and moats, but valuations may have partially reflected expectations; data layer projects (PrismaX) have high elasticity but come with data quality risks; financial layer projects (Virtuals) and application layer projects (XMAQUINA) are more dependent on ecosystem prosperity and community activity.

Reflecting on the mobile industry's evolution from the "shanzhai phone era" to "Android/iOS dual dominance," the Web3 robot track may be experiencing a similar early stage. As OpenMind's founder said, there will not be only one winner in the future; many robust participants will emerge. For investors focusing on this track, now is the window period for observation, learning, and selective participation. The future of machines becoming on-chain economic entities is transitioning from science fiction to reality.