This edition of the newsletter dives into the concept of DePAI (Decentralised Physical AI) which represents the convergence of AI x robotics x crypto - creating a framework where autonomous machines can operate with sovereignty through crypto infrastructure and token incentives. While this emerging ecosystem promises democratised ownership and aligned incentives across robotic systems, it faces significant challenges including insufficient blockchain infra, safety risks, regulatory scrutiny, and the potential for speculative investment to prioritise hype over technical fundamentals. We'll also share some interesting articles, portfolio updates and market highlights.
1. Research Articles
a) Solana vs Base: The Battle For Stablecoin Payments
• The article compares Solana and Base blockchains as platforms for stablecoin payments, examining their performance, fees, market demand, and regulatory status. It concludes that while Base has advantages through its Coinbase connection and US presence, Solana may be better positioned for long-term stablecoin payments due to its neutrality, diverse ecosystem, and resistance to censorship.
• Solana --> Decentralised, high demand, diverse ecosystem, censorship resistance; Base --> Centralised, strong US presence, Coinbase/USDC advantages, rapid development.
Solana vs Base: The Battle For Stablecoin Payments - Squads Blog
Companies and startups are increasingly turning to Solana and Base to overhaul their payments infastructure.
b) Make Revenue Great Again
• As crypto market liquidity tightens, founders are shifting focus from speculative token value to sustainable revenue models, with many adopting transaction-based fees similar to Web2 companies like Stripe and Shopify
• Several crypto projects (including Sky, Ronin, Jito, Kaito, and Gearbox) are implementing token buyback mechanisms to tie token value to protocol revenue, mirroring corporate stock buybacks in traditional markets.
Make Revenue Great Again
Web3 native business models and revenue generation
c) The path to secure and efficient zkVMs: How to track progress
• Security Challenges Require Years of Formal Verification: Despite industry narratives suggesting immediate deployment readiness, zkVMs face significant security vulnerabilities that will take 2+ years to establish correct protocols and 4+ years to achieve fully verified implementations. The security roadmap includes formally verifying all system components (from proof systems to constraint systems to implementations), with some deployments currently representing "an expensive way of pretending a system is secured by a SNARK."
• Performance Limitations Make Real-World Deployment Impractical: Current zkVMs operate with overhead approaching one million times the cost of native execution. The author establishes concrete performance milestones: reducing overhead to 100,000x native execution (Speed Stage 1), then 10,000x (Speed Stage 2), while solving memory constraints (under 2GB for Memory Stage 1, under 200MB for Memory Stage 2). While pre-compiles offer temporary optimizations, they introduce security risks and don't address fundamental inefficiencies in the core proof systems.
The path to secure and efficient zkVMs: How to track progress
The stages for zkVM development, the completion of which will take several years. Don't let anyone tell you differently.
2. Portfolio Highlights
a) Soneium
• Soneium partners with Line to accelerate on-chain gaming through integration with LINE MINI-App
• Builders can leverage on Soneium's support—combining community building, strategic marketing, and IP collaboration
Soneium | Soneium and LINE MINI App: Expanding web3 Globally
Soneium, powered by Sony Block Solutions Labs, is collaborating with LINE to bring four successful mini-apps onchain in coming months. This collaboration opens up new opportunities for web3 adoption. With approximately 200 million active users already engaging with LINE, integrating its mini-apps into the Soneium ecosystem enhances user experience, bringing smoother, more scalable, and accessible blockchain interactions.
b) Particle Network
• "$PARTI Diamonds" --> a seasonal rewards program for traders on the UniversalX platform. Users earn it by trading actively on UniversalX + referrals
• UniversalX Platform: "the first chain-agnostic trading platform," allowing users to trade across different blockchains without bridging
Time to Trade on UniversalX: Introducing $PARTI Diamonds
We've come a long way. First, we put chain abstraction on the map and built a community around it. Then, we released the first-ever chain-agnostic dApp-all within a few months. Now, we're thrilled to introduce you to $PARTI Diamonds. Community OGs and pioneering traders of the world-
c) Sonic
• Sonic's Fee Monetization (FeeM) system revolutionizes blockchain economics by allowing developers to earn up to 90% of the network fees their applications generate, creating a direct revenue stream similar to how YouTube pays creators but with transparent, permissionless payouts. This aligns incentives between developers and the network, transforming transaction costs into sustainable income and shifting the focus from short-term hype to long-term value creation, while platforms like Beets are already leveraging FeeM to create self-reinforcing growth loops through buybacks and community-directed incentives.
• Sonic x Royco Partnership - Royco will distribute ~200m $S tokens as part of their Szn 1 incentive campaign. Users can earn 3 types of incentives: App Gems (redeemable for $S), Royco Bonus Gems (exclusive to Royco users), and Sonic Activity Points
173
3. DePAI: The Convergence of Decentralised Networks and Robotics
Introduction
As the market resets itself after the AI agent frenzy of the past 5 months, the question on everyone's mind is what comes next? Today we explore the transition from agentic AI to physical AI (robotics), where a new narrative has been formed combining robotics with crypto incentives termed as DePai (decentralised physical AI), a spin on DePin (decentralised physical infrastructure networks). In this article we give a recap on DePin, an introduction to robotics and how it can be combined with crypto, as well as some of the risks
DePin basically leverages on token incentives to build and operate real-world infrastructure and incentivise a large network of operators to contribute to the ecosystem. These networks are economic models where users can contribute and benefit from decentralised systems including wireless networks (Helium), data storage (Filecoin) and computing resources (io.net). Emerging from the DePin narrative (that has since dwindled) is the concept of Decentralised Physical AI (DePAI), a term that gained prominence following NVIDIA CEO Jensen Huang's introduction of "Physical AI" at CES in January 2025. It can be used to describe the intersection of artificial intelligence (AI), robotics, and crypto/blockchain technology. In short, DePai not just decentralises the computational power supporting AI models, but also reimagines how machines operate in the physical world through crypto.
At its core, DePAI envisions robots and autonomous machines operating not just independently but with agency and sovereignty on crypto infra. This vision extends far beyond mere automation to create systems where embodied AI can perceive, interact with, and thrive in the real world while operating within decentralized economic frameworks (in most cases, with the usage of tokens). Consider the trajectory of AI development - from language-based generative AI focused primarily on content creation to agentic systems capable of independent decision-making and real-world interaction, now with robots that can complete real life tasks as well.
The DePAI ecosystem consists of 7 interconnected layers that collectively enable this new paradigm (inspired by a post from Peaq)
Hardware layer ("the body): the foundation, consisting of physical robots that allow AI models to interact with their environment. Also includes cameras, lidar, magnets, sensors and motors
Software layer ("the brain"): agentic AI models that give these robots autonomy and decision-making capabilities. Also includes a mix of semiconductors and software for GenAI models
Data layer: provides the essential information AI requires to learn about and navigate the real world
Spatial intelligence enables robots to comprehend and interact with their physical surroundings in real-time
Infrastructure networks deliver the essential resources (storage, compute, energy) needed for ongoing operation
The Machine Economy layer: integrates all these components with protocols that ensure interoperability and coordination
Organizational structures like DePAI DAOs allow individuals, communities, and businesses to collectively own and benefit from these technologies
Robotics
Most of the robotics companies have so far focused on industrial robotics development which are primarily designed for manufacturing environments, performing repetitive and mundane tasks with better precision than humans. For decades, robotics remained largely confined to structured industrial settings where environmental variables could be tightly controlled. According to a report by Morgan Stanely, 73% of companies confirmed to be involved in humanoids are based in Asia (56% in China), highlighting Asia's dominance in the sector. Chinese companies are showing the most impressive progress, benefiting from established supply chains, local adoption opportunities, and government support.
The robotics landscape has drastically changed in recent years, where modern robots are not just developed for specialised industrial machines but to increasingly versatile consumer products, autonomous vehicles and even humanoid platforms. The leading companies include Figure, 1X, Clone Robotics, Boston Dynamics, Tesla and Agility Robotics. The field has been transformed by advances in materials science, battery technology, sensor arrays, computer vision, and AI. For AI, LLMs especially has enabled robots to interact with unstructured environments in ways previously impossible, recognizing objects, planning paths, and making decisions based on real-time data.
The key point is studying the shift from simply robots that automate predefined sequences to creating systems that can learn, adapt, and operate with increasing degrees of autonomy through AI and LLMs. GenAI is transforming robotics by allowing machines to learn through natural language, imitation, and simulation. This progression is similar to the evolution in AI from purely generative models to agentic systems capable of proactive decision-making and to give an example:
Generative AI might create a grocery list based on your input
Agentic AI analyses your health data, creates optimized meal plans, suggests recipes, monitors your refrigerator inventory, and automatically orders groceries when supplies run low
In the case of crypto, this evolution is similar to that of DeFai (Decentralised Finance AI) where agents create optimized yield sourcing capabilities and automatically generates yield strategies according to the preferences of the user. Just like AI language models that got smarter with more data and computing power (from just text output to AIs capable of actually completing tasks), robots might also follow the same path. The only difference is AI operates online, compared to physical robots operating in real world with endless variations and unpredictability compared to the digital world. Skeptics auestion if they can really master using the same methods that work for language - text follows rules and patterns, while reality is more complex and random.
Robotics x Crypto
From crypto gaming to DeFi, anything with incentives and tokens gives rise to an internal economy with certain flywheels. For robotics, this represents a decentralized framework where autonomous machines interact, transact, and coordinate without requiring centralized intermediaries and uses a distributed network that is governed by a token. Tokens allow for better coordination, resource allocation and value creation.
The token integration comes in at the infrastructure level, where DePIN networks provide essential resources for robotic systems. Companies leverage token incentives to ensure a wide network of distributed resources which also reduces bottlenecks. Another advantage is the democratisation of robotics development and ownership. Traditional robotics companies require substantial capital investment, limiting participation to well-funded corporations or venture-backed startups. Crypto mechanisms enable fractional ownership models where individuals can collectively invest in robotic assets through DAOs or community fundraise platforms.
The last aspect of is the integration of token incentives, where users can complete quests and earn rewards in exchange for collecting valuable real-world data. By transforming data collection into a gamified experience with economic incentives, they created a model that simultaneously entertains users, rewards participation, and advances robotic capabilities. The token economics create aligned incentives across the entire ecosystem: robot owners earn returns by making their hardware available for use —> data contributors receive compensation for the information they provide (similar to DePin plays) —> developers gain access to distributed testing environments while retail users benefit from improved services —> token holders participate in the overall growth of the network in terms of governance.
This also creates a strong flywheel for projects: gamification attracts users —> refer their friends —> earn points and more rewards as incentives—> more attention on the product and user base grows which improves the service —> demand increases --> more revenue --> team able to raise more funds to expand and grow —> more capital for marketing and ensuring the token does well —> token price increases --> attract even more users and speculators.
Some of the associated risks
Current L1 infra probably is insufficient to support the deployment of autonomous robotic systems, recently Peaq launched their mainnet to specialise on DePin economic systems and the machine economy. While some DePINs provide necessary data and computational resources, the real-time processing requirements of physical AI far exceed what existing networks can consistently deliver. This challenge becomes obvious for time-sensitive dApps like autonomous navigation in dynamic environments, where decisional latency could have severe consequences.
Combining the internet financial system with real life physical activities brings up a unique set of challenges, where physical AI carries inherent safety risks that require careful oversight. Determining who bears responsibility when autonomous systems cause harm, how safety standards should be established and enforced, and what recourse affected parties have requires sophisticated governance frameworks that balance innovation with accountability. Who determines this governance frameworks - token holders or the foundation? The decentralized nature of DePAI systems complicates these questions by distributing authority across diverse stakeholders without clear hierarchies of responsibility.
The speculative nature of many crypto projects presents additional risks for DePAI development. If investment is driven primarily by token speculation rather than fundamental utility, resources may flow disproportionately to projects with compelling narratives rather than those addressing critical technical challenges. We have seen this played out a few times this cycle, across AI, AI Agents and DeSci as well - when tokens do not have a good use case and is only meant to be speculated. Once the token dumps, teams run of of runway to continue building the project.
Conclusion
The convergence of decentralized networks and robotics through the DePAI framework represents a potential paradigm shift in how autonomous systems are developed, deployed, and governed and could represent the next iteration after DePin and autonomous agents. By combining the physical capabilities of robots with the economic mechanisms of crypto and the intelligence of advanced AI, DePAI creates possibilities for more accessible, versatile, and equitably distributed robotic systems. Rather than concentrating ownership and control of these transformative technologies among a few corporations, the decentralized approach enables broader participation in both the development and benefits of robotic automation. Robotics is one of the most significant technological and economical developments of the coming decades, with substantial implications for labour markets and manufacturing.
As with many emerging tech and narratives in crypto, the ultimate impact of DePAI will depend not just on what is technically possible but on how these capabilities are deployed and governed and the balance between the project and value accrured back to token holders. It is necessary to consider the benefits, where the decentralized nature of DePAI could prove to be its greatest strength - enabling diverse perspectives and priorities to shape the evolution of robotic systems in ways that reflect broader human values and needs rather than narrow commercial imperatives.
*Disclosure: All information provided in this newsletter is for general informational purposes only and does not constitute professional nor investment advice.
