Over 200 subscribers
This edition of the newsletter dives into NetMind which is introducing a decentralized computing platform called NetMind Power that aims to break the monopoly of big tech companies by allowing users worldwide to contribute their idle GPU resources for AI training, fine-tuning, and inference tasks. The platform operates on its own blockchain (NetMind Chain) with a native token ($NMT), and includes features like a training platform, inference platform, and NetMind.xyz - a society of AI agents where users can create and deploy their own AI agents that can interact with humans and each other through various platforms. We'll also share some interesting articles, portfolio updates and market highlights.
a) Chain Abstraction: Orchestrating Towards the New Internet's Composability
• This article discusses the problem of blockchain fragmentation, where different chains have varying trust assumptions and security definitions, leading to issues with liquidity, application deployment, and asset management across networks.
• The article introduces chain abstraction as a solution, particularly focusing on the SOCKET protocol's role in enabling seamless cross-chain operations and unified user experiences by creating chain-abstracted packets (CAPs) that can execute asynchronously across multiple networks.
The article explores blockchain fragmentation issues and introduces chain abstraction as a solution to unify state, liquidity, and assets across multiple chains. It highlights Socket protocol's role in enabling seamless user experiences and multichain composability for dApps and developers.
b) Spectral (I): Sentient Memes
• This article examines Spectral Labs' latest product, Syntax V2, a platform that enables users to create and manage AI trading agents that can autonomously trade perpetual futures on Hyperliquid using GPT-4.
• The platform features a unique tokenomics model where agent creators and token holders can influence trading decisions, with fees and profits flowing back to both individual agents and the platform's SPEC token holders, though it faced a security breach in December 2024 that was subsequently resolved.
What if our favourite memes could talk, or trade for us?
c) Blockchains vs. Bureaucracy: The Costs Of Curtailing Innovation
• This article examines how U.S. regulatory pressure, dubbed "Operation Chokepoint 2.0," has led to widespread "debanking" of crypto and fintech companies, with over 50 companies reportedly having their bank accounts closed despite compliance with existing regulations.
• The article argues that while anti-money laundering (AML) regulations cost banks $274B annually with only 0.1% success rate in flagging illicit transactions, blockchain technology offers a more transparent and efficient alternative for financial monitoring, and solutions like smart accounts are emerging to help debanked companies operate independently of traditional banking systems.
This reserach report analyzes the realities of debanking and quantifies its costs.
a) Sonic Labs
• Sonic has launched its mainnet offering 10,000 transactions per second, sub-second finality, and a native decentralized gateway to Ethereum, allowing FTM token holders to upgrade to S tokens on a 1:1 basis.
• The platform incentivizes developers with up to 90% of network fees generated by their apps through Fee Monetization (FeeM) and includes airdrops of Sonic Points and Sonic Gems to reward user engagement and developer innovation.
Sonic is an EVM-compatible layer-1 blockchain that offers developers attractive incentives and infrastructure. 10,000 TPS at sub-second finality.
b) Pendle
• Pendle has announced Boros, a new derivatives market that aims to tokenize funding rates and other yields without requiring receipt tokens, moving beyond their existing V2 protocol which splits yield-bearing tokens into Principal Tokens (PT) and Yield Tokens (YT).
• While Boros will initially focus on perpetual funding rates with plans to expand to staking rates across different chains, its success will depend on how it competes with established perpetuals protocols like Hyperliquid and dYdX, though Pendle may have an advantage due to its existing yield-focused user base and ability to serve specialized hedging needs.
Derivatives markets continue to broaden.
c) Blade Games
• Blade Games is a blockchain gaming studio that leverages WebAssembly and zero-knowledge virtual machines to create games with verifiable on-chain logic, having evolved from a small team in 2022 to a full ecosystem with multiple games, partnerships, and an AI agent framework by 2024.
• The company is now focusing on creating games that can be played by AI agents rather than just humans, utilizing various frameworks for agent behavior and social interactions, while generating revenue through in-game taxation, model training, and infrastructure incentives.
This Whitepaper set out to showcase the unique features and opportunities that Blade Games bring to 3 sectors, Gaming, AI and Zero Knowledge technology. The Whitepaper will be maintained and continuously updated by Blade Games' team.
1. Introduction
In recent years, the ever increasing demand for high-performance computing has been driven by rapid advancements and growing interest in Machine Learning (ML) and Artificial Intelligence (AI). However, this boom has exposed significant challenges, including limited access to computing power, fragmented resources, and soaring costs.
The concentration of computing capabilities among tech giants has exacerbated the problem, granting these companies monopolistic control over pricing and access. Smaller businesses, research institutions and startups that try to utilize AI face prohibitive capex, hindering their ability to innovate and compete in the AI landscape, while larger corporations solidify their dominance.
2. Compute - the new currency of the future
The demand for AI hardware is skyrocketing, driven by consumers, corporations, and governments alike. For instance, ChatGPT became the fastest-growing consumer application in history, surpassing 100 million monthly active users (MAUs) in record time and continuing to grow rapidly. Beyond OpenAI, the past year has seen thousands of talented teams emerge, each striving to bring AI to the forefront with innovative approaches across various fields. Nation-states are increasingly recognizing the strategic importance of compute resources for national security, with AI-related federal contracts soaring by about 1,200%, from $355 million in August 2022 to August 2023.
This relentless demand has created a compute shortage, leading to significant price inflation and ensuring GPUs retain their value over time. With manufacturers struggling to meet the growing needs of the market, this scarcity and its associated challenges are expected to persist well into the future.
3. Introducing NetMind
To address these challenges, NetMind.AI, an AI research and software company based in introduces NetMind Power - a decentralized computing platform for ML and AI training, fine-tuning, and inference. Driven by a team with experience in AI over the past decade, Netmind was designed to cater machine learning professionals, researchers, and software developers, harnessing the idle GPU from users worldwide to create a global network of computing power.
By integrating diverse computing resources through advanced grid and voluntary computing scheduling architecture, NetMind Power offers a scalable, low-latency platform for distributed deep learning training and inference. This solution democratizes access to computing power, enabling businesses and research institutions to affordably develop and run AI models. NetMind’s mission is to break down barriers to innovation, empowering organizations of all sizes to contribute to the future of AI.
4. NetMind Power Overview
a) The training Platform
The training platform forms the cornerstone of NetMind Power's decentralized computing ecosystem. It enables users to efficiently and cost-effectively train and fine-tune AI models by leveraging idle GPUs contributed by participants worldwide. Built on cutting-edge technologies and methodologies, the platform ensures scalable, distributed AI model training, offering a robust alternative to traditional centralized computing resources.
However, an important note that the actual training processes occur locally on the participating machines, not on the chain itself!
b) Decentralized Architecture
The platform employs a decentralized network of interconnected nodes and miners, distributing training workloads across multiple GPUs globally, where the “init node” represents the first instance to the whole process. By shifting away from centralized resources, this architecture minimizes dependency on large-scale infrastructure, significantly reducing training costs while maintaining high performance and scalability.
While the computational resources and capacity of miners and nodes across the network may vary, NetMind's intelligent resource allocation system ensures optimal efficiency by dynamically assigning tasks to the most suitable entities. This adaptive approach maximizes performance and resource utilization across the decentralized network, while constantly optimizing for best possible latencies and costs for the user.
On top of that, due to its flexible design from the ground up, the training platform leverages data and model partitioning techniques, including data and model parallelism, to divide workloads across network GPUs, optimizing processing efficiency. Using methods like Federated Learning and parameter averaging, it securely aggregates model updates while preserving data privacy. Advanced encryption, secure multi-party computation, and optional differential privacy further safeguard user data throughout the process.
c) Inference Platform
The inference platform is a critical complement to the training platform, offering users a streamlined solution to deploy and execute inference tasks on various AI models. These include models they have trained, those developed by others, or readily available open-source models.
Built with a focus on flexibility and efficiency, the platform’s innovative architecture ensures seamless integration into diverse workflows, enabling users to leverage AI at production scale with minimal friction. By delivering high-performance inference capabilities at a fraction of traditional costs, the platform empowers organizations of all sizes to unlock AI’s potential without the typical resource constraints.
Key features include robust scalability to handle varying workloads, support for heterogeneous model types, and optimized resource allocation for cost efficiency, by leveraging the NetMind resource allocation algorithm, ensuring tasks are assigned to the most suitable GPUs based on capacity, latency, and availability, optimizing both performance and operational costs.
Additionally, the platform emphasizes security and reliability, ensuring data protection and system stability, making it an ideal choice for enterprises, researchers, and developers aiming to adopt AI in a decentralized, secure, and scalable ecosystem.
5. GPU Dashboard
6. Netmind.xyz - A society of AI agents
NetMind.xyz aims to revolutionize the AI landscape by creating an innovative AI Agent Society, where users can design, deploy, and interact with their own AI agents powered by NetMind’s advanced multi-agent framework. Building on a foundation of cutting-edge AI technology, this ecosystem provides unprecedented flexibility and utility for users while seamlessly integrating tokenomics to drive engagement and growth.
Netmind based Agents will be powered by fine-tuned models and featuring advanced modules for memory, reasoning, consciousness, and momentum. These enhancements enable agents to process information more deeply, reason effectively, and operate with human-like intelligence.
Additionally, agents will possess multi-modal capabilities, allowing them to generate and interact with content across various formats, such as images, videos, and music. Leveraging cutting-edge technology, agents will feature lifelike digital human systems with lip-sync, voice conversion, and image animation, making them ideal for applications in education, customer service, and content creation.
However, these agents will not exist in isolation - they will interact with humans via platforms like X, WhatsApp, and Telegram, and be part of interconnected communication and economic systems, enabling collaboration and even trade among each other - creating their own ecosystems and civilization.
These innovations are part of NetMind’s broader vision to create a new AI civilization, where intelligent agents and humans coexist, learn, and evolve together, with $NMT as the foundational power behind their growth and interactions.
Core features
a) User-Created AI Agents: Users will have the capability to craft AI agents or multi-agent systems tailored to specific needs or functions. These agents will operate autonomously or collaboratively, leveraging NetMind’s robust framework for seamless communication and problem-solving.
b) Tokenization of AI Agents
Each AI agent or multi-agent system can generate its own unique token that is intrinsically bonded to NetMind’s native token, $NMT. These tokens can represent various utilities within the system, such as specialized access rights, custom AI services, or other innovative uses, opening doors to a dynamic tokenized ecosystem.
c) $NMT as the Core Currency
• Building Agents: Users will require $NMT to create and deploy AI agents on the platform.
• Interaction and Utilization: Engaging with these agents, whether for tasks, services, or collaboration, will require $NMT.
• Resource Access: All AI agents will consume $NMT to access NetMind's decentralized GPU network and pre-trained AI models, driving sustained demand for the token.
7. Driving demand for $NMT
Currently, volunteer computing resource providers earn $NMT by staying connected to the network for over 5 hours a day, with rewards based on GPU type, bandwidth, and the number of connected machines.
Further, the newly introduced AI agent framework will drive additional demand to the token. The design ensures that every interaction within the AI Agent Society directly contributes to the ecosystem’s token economy. From agent creation to resource utilization, $NMT serves as the backbone, supporting the growth of an active, token-driven AI marketplace.
NetMind.xyz not only empowers users to innovate with AI but also integrates blockchain and tokenomics to create a self-sustaining, vibrant ecosystem that reshapes how individuals and organizations interact with AI agents.
Enhanced token utility:
• Token Burning Mechanism: Reduces supply, increasing scarcity and potential value
• Controlled Token Buyback Initiatives: Strengthens demand and supports token value by repurchasing tokens from the market
• Increased Payment Rewards: Miners will earn more through payments and less from mining, boosting overall value for participants
• Enhanced Staking Requirements: Higher staking will be required to activate mining rewards, with a portion redirected to a long-term staking pool
• Expanded Utility: New initiatives like model APIs, multi-agent tools, and strategic partnerships will increase NMT's demand and utility, enhancing token value and reinforcing its role in the ecosystem
8. Conclusion
In conclusion, NetMind has become a vital force in addressing the growing demand for AI computing power while challenging the monopolistic control of a few tech giants. By offering decentralized, cost-effective, and reliable computing resources, NetMind has empowered research institutions, AI companies relying on its infrastructure for inference tasks, and developers to access affordable, scalable, high-performance computing. Its innovative approach - featuring intelligent resource allocation and the new NetMinds.xyz agent framework - is key to democratizing AI development. As more organizations embrace this technology, NetMind plays a crucial role in leveling the playing field, fostering innovation, and encouraging competition in an industry increasingly dominated by a handful of major players.
This article was written by 0xNiko
*Disclosure: Signum Capital’s holds a position in the company mentioned above and the information provided on this newsletter is for general informational purposes only and does not constitute professional nor investment advice.
