The iteration speed of Artificial Intelligence (AI) is beyond imagination. From the early perception-based AI that could only understand images and text, to generative AI focused on content creation, and now to intelligent agents capable of independent decision-making, AI has come a long way. Earlier this year, NVIDIA CEO Jensen Huang introduced the concept of Physical AI at CES (Consumer Electronics Show), which refers to AI that can run, reason, plan, and act. This also sparked discussions in the industry about DePAI (Decentralized Physical AI). Now, another new term, DeSPIN, is gradually entering the public’s view.

DeSPIN, short for Decentralized Spatial Intelligence Network, combines spatial intelligence technology with the decentralized principles of blockchain and Web3. The relationship between AI and the real physical world is redefined through methods such as 3D environment modeling and real-time spatial collaboration. This article will delve into the DeSPIN track, exploring concepts, core operating principles, potential challenges, and highlighting some noteworthy projects.

What is DeSPIN?

Spatial Intelligence is an integrated capability that combines various technologies and methods to perceive, understand, and interact with 3D physical space, enabling systems to make real-time decisions and actions. Its core goal is not just to “see” the world, but to empower systems with autonomous cognitive abilities to comprehend and operate within the physical world.

Based on this concept, DeSPIN can be seen as a distributed system focused on 3D spatial perception and collaboration, where users are rewarded for contributing spatial data to co-build an open and fair shared ecosystem. Powered by core technologies like computer vision, sensor technology, and machine learning, DeSPIN is widely applied in smart city planning, autonomous driving, gaming, environmental monitoring, and natural resource management, offering notable benefits in cost reduction and resource optimization.

DeSPIN has a close collaborative relationship with AI, DePAI, and DePIN (Decentralized Physical Infrastructure Network). To some extent, DeSPIN relies on the hardware and spatial data provided by DePIN, is driven by AI, and serves as a key component of DePAI.

Core Operating Principles

The core of spatial intelligence lies in building a “world model”—an environment simulation system constructed from raw data. Through complex processing and analysis, it enables comprehensive understanding and prediction of the external world, allowing for more reasonable and optimized decision-making in various contexts.

In summary, the core operating principles of spatial intelligence include:

• Spatial Perception: Capturing spatial information through sensors (like cameras, GPS) and devices (such as smartphones, dashcams) to collect environmental data.
• Data Processing: Integrating and analyzing data using technologies like machine learning and computer vision to build 3D spatial models.
• Knowledge Representation: Converting data into semantic spatial information (e.g., obstacle locations, buildings, vegetation) to form an interactive world model.
• Decision Support: Using the model to predict environmental changes and generate optimized plans (such as autonomous driving route planning or natural disaster risk assessment).

Within the DeSPIN ecosystem, users earn token rewards that can be monetized by contributing real-world data through various network participation methods such as Sky-to-Earn (flying), Drive-to-Earn (driving), or Train-to-Earn (training).

Track Project Overview

Although the DeSPIN track is still in its early stages, several forward-looking pioneers have quietly begun laying the groundwork, aiming to gain an early edge in this emerging space. Below are some noteworthy projects worth paying attention to.

• NATIX Network

NATIX is a global smart camera mapping network that builds a decentralized, dynamic world map to support autonomous driving, physical AI, and more. In this network, users earn rewards by using their cameras to map their surrounding environments. According to the official data, NATIX has over 243k registered drivers across 171 countries and regions, collectively mapping more than 153 million kilometers.

(Source: natix.network)

NATIX was founded in 2022 and completed a $3.5 million seed round in 2023 and a $4.6 million strategic round in 2024. Investors include Blockchange, XYO, Borderless Capital, Tioga Capital, Inception Capital, and Laser Digital.

NATIX has two core products: the Drive& app for driver assistance and the VX360 plug-in device, both of which offer user rewards. Drive& supports both iOS and Android. Users can earn through Drive-to-Earn activities, with rewards in $NATIX tokens, products, and discounts. Currently, Drive& provides dashcam functionality, with features like speed alerts, navigation assistance, and parking availability to be added later. VX360 is a driving device currently available for pre-order at $279.99. Users can plug it into a Tesla and use the companion app to earn mileage-based rewards.

$NATIX is the core token of the NATIX ecosystem and launched in July 2024. With a total supply of 10 billion tokens, it circulates on the Solana network. $NATIX serves multiple purposes within the ecosystem, such as accessing platform features, staking for rewards, and participating in governance.

• DIMO

DIMO is a decentralized vehicle data insights network focused on unlocking the value of car data while enabling users to earn rewards through Drive-to-Earn. According to the official site, DIMO currently supports over 50 unique car models and more than 173,000 vehicles.

(Source: dimo.co)

At this stage, users manage their vehicles via the DIMO app, earning rewards while easily accessing data such as vehicle health, mileage, and GPS location. To enhance the user experience, DIMO has developed two proprietary plug-in devices, priced at $782 and $1,335 respectively, which help users accelerate their earnings. Looking ahead, the team plans to expand use cases for car data, including AI-powered auto repair and smart parking solutions. DIMO also aims to extend beyond vehicles to include non-mobile IoT applications.

DIMO raised $9 million in 2022 and $11.5 million in 2024, with investors including CoinFund, Variant Fund, Slow Ventures, and Borderless Capital. $DIMO is the network’s native token, launched in December 2022 with a total supply of 1 billion. Of that, 45% is allocated to user incentives, distributed over a 40-year period. In the first year, 1.105 million $DIMO tokens are distributed weekly, with an annual reduction of 15%. $DIMO plays a central role in the ecosystem, including access to services, staking rewards, and participation in governance.

(Source: docs.dimo.org)

• Hivemapper

Hivemapper aims to build a new map for humanity by allowing users to collect high-quality street-level imagery using dashcams, earning $HONEY tokens in return. According to the official data, the Hivemapper map now covers 31% of the global area, with over 476 million kilometers of roads mapped.

(Source: beemaps.com)

In April 2022, Hivemapper raised $18 million in a round led by Multicoin Capital, with participation from Solana co-founders Anatoly Yakovenko and Raj Gokal, Helium founder Amir Haleem, and other industry leaders.

To support high-quality data collection, Hivemapper has developed several proprietary dashcams, with Bee as the flagship product, currently priced at $589. Hivemapper also offers an AI-powered fleet management solution called Beekeeper, which supports real-time vehicle monitoring, trip tracking, and automatic event detection. Upcoming features include AI-based smart route planning. In addition to driving, users can participate as AI Trainers, earning tokens by completing tasks such as labeling and editing data to improve the Map AI.

$HONEY is the core token of the Hivemapper network, issued on Solana and launched in December 2022. It has a capped supply of 10 billion and uses a deflationary mechanism by burning tokens when network services are accessed. 40% of the total supply is allocated to user rewards, with distribution amounts based on the progress of map creation.

• Spexi

Spexi is the first solution designed for crowdsourced and standardized aerial imagery capture, focusing on Fly-to-Earn. It enables drones to earn rewards by flying and capturing large-scale aerial images, supporting use cases like smart cities and natural resource monitoring. According to the team, Spexi has completed full imaging across 160+ cities globally, recorded over 58,000 on-chain transactions, and collected 3.7 million ultra-high-resolution aerial images, with nearly 30 enterprise clients.

(Source: docs.spexi.com)

Spexi supports drones such as DJI Mini 2, Mini 3, and Mini 3 PRO. Users must download a dedicated app and comply with local regulations to capture images. To standardize the capture process, Spexi uses hexagonal grid units called Spexigons. Different types of Spexigons have different reward requirements for pilots. Spexi rewards users based on a Proof-of-Capture (POC) mechanism.

Currently, Spexi has not launched a token. Users are rewarded in-app with cash and points. According to the roadmap, the mainnet is scheduled to go live in Q2 this year. Spexi has raised a total of $17 million in funding across 2022 and 2024, with investors including Blockchange, Protocol Labs, Alliance DAO, Dapper Labs, FJ Labs, and Moonshots Capital.

• GEODNET

GEODNET is a decentralized precision positioning network that uses Real-Time Kinematic (RTK) navigation to enhance satellite data (from systems like GPS, GLONASS, Galileo, or BeiDou) for centimeter-level accuracy, powering applications like robotics, drones, and autonomous vehicles. As of March 31, the network has attracted over 14,000 miners across 142 countries, with RTK usage exceeding 149,000 GB.

GEODNET supports multiple base stations, such as Hyfix’s MobileCM triple-band GNSS station. Users can earn $GEOD rewards by contributing data through hosting these stations. The network has also released a vehicle navigation device called GEO-PULSE, which offers 10x the accuracy of GPS and is priced at $149. Users can earn $GEOD for completing specific navigation tasks.

$GEOD is the core token of the GEODNET ecosystem, launched in August 2022 and available on Solana, Polygon, and IoTex. With a maximum supply of 1 billion, 35% is allocated for miner rewards, and the base reward halves annually. $GEOD serves both utility and governance roles, including payment for RTK data services and participating in ecosystem decisions.

Notably, $GEOD follows a deflationary model, with 80% of data revenue used to buy back and burn tokens. According to Dune data, as of March 31, 137 million $GEOD have been distributed as mining rewards, and 12 million $GEOD (worth around $1.89 million) have been burned.

(Source: Dune@geodnet_console)

Over the past few years, GEODNET has raised over $15 million in funding from notable investors including Borderless Capital, IoTeX, North Island Ventures, Pantera Capital, Animoca Brands, and Multicoin Capital.

• Auki Network

Auki is building a decentralized machine perception network and collaborative spatial computing protocol. By integrating with devices, Auki forms trustless clusters that enable the exchange of data and computing resources. Users who contribute services and resources to the network are rewarded. According to the official site, Auki currently operates 727 relay servers and 631 domain servers globally.

(Source: dashboard.auki.network)

$AUKI is the native token of the Auki Network, launched in August 2024 and circulating on Base. The total supply is capped at 10 billion tokens, with approximately 1.625 billion currently in circulation. $AUKI features a burn mechanism until the token supply is halved. As the core token of the ecosystem, $AUKI serves multiple purposes: accessing services through token burning, staking to become a node, and earning additional rewards.

• Alaya AI

Alaya AI is an open, composable Web3 data sampling and auto-labeling AI infrastructure that allows anyone to contribute AI data and complete AI training tasks in a gamified way to earn rewards. According to the official site, Alaya AI has over 3.62 million registered users and more than 300,000 daily on-chain transactions.

(Source: @Alaya_AI)

Alaya AI is set to launch its data annotation toolset, Alaya AI Pro, which provides pixel-level image annotation and segmentation. It supports real-time processing of both static and dynamic visual data, making it suitable for applications in autonomous driving, robotics, and the Internet of Things (IoT).

In the Alaya AI ecosystem, $AGT is the core asset, with a maximum supply of 5 billion tokens, and the Token Generation Event (TGE) is forthcoming. $AGT serves both utility and governance functions, including in-game consumption, access to ecosystem services, advanced task bounties, and governance voting. In addition to $AGT, NFTs are important assets within the Alaya AI platform, with different varieties and rarities affecting user platform access and reward amounts. To strengthen community incentives, the team plans to use platform revenue to buy back $AGT and redistribute it as user rewards.

Potential Risk Analysis

DeSPIN is still in the concept validation stage, but it has already shown immense potential in building the “data democratization” infrastructure of the physical world. As a space data-driven intelligent system with high autonomy and environmental adaptability, DeSPIN is ushering in a new era full of opportunities and challenges.

However, DeSPIN faces multiple challenges on the road to the future:

• Regarding technology, the realization of spatial intelligence depends on integrating several cutting-edge technologies, including 3D generation, multimodal large models, and machine learning, all of which come with significant uncertainty and integration complexity.
• Regarding costs, spatial understanding requires substantial data and powerful computational and storage resources, which can be a high threshold for small and medium-sized teams. From the user’s perspective, it is essential to consider the hardware investment’s payback period or develop reasonable strategies to improve profit efficiency.
• In terms of data, ensuring the privacy, security, and authenticity of data remains a long-term challenge.
• The token economic model is crucial for the ecosystem’s development in terms of user incentives. However, reward-based mechanisms like Move-to-Earn and Learn-to-Earn have struggled to achieve long-term growth in previous attempts. The team must carefully design and balance values and mechanisms to maintain the ecosystem’s dynamic sustainability.

Conclusion

With the continuous integration of Web3 and AI technologies, DeSPIN may become the super interface that connects the digital and physical worlds, driving more efficient collaboration, decision-making, and interaction across various fields. However, the road from concept to practical application is often full of twists and turns, with numerous uncertainties and challenges. Investors must always maintain a rational perspective.-