This article will focus on projects in this field that have not yet issued coins, in order to find early Alpha.
Written by: TechFlow TechFlow
AI has become the main line of today's technology circle, and the integration of AI and cryptocurrency has become the focus of capital pursuit. Autonolas (token Olas), an AI-based oracle, has increased by 559.3% in the past month. Tau (token AGRS), a public chain focusing on AI and has its own logic programming language, has increased by an astonishing 2560.0 in the past month. %.
These numbers not only represent the prosperity of the market, but also reflect the potential opportunities at the intersection of AI and cryptocurrency. Perhaps this sector will become a new gold mining mecca for investors.
This article will focus on projects in this field that have not yet issued coins, in order to find early Alpha.
1.io.net: Decentralized GPU computing power provided based on Solana
io.net is a decentralized computing network that supports the development, execution and scaling of ML (machine learning) applications on the Solana blockchain, leveraging the world's largest GPU cluster to allow machine learning engineers to perform the equivalent of centralized services Access distributed cloud service computing power at a fraction of the cost.
Existing centralized cloud computing services have the following problems for engineers:
- Availability is limited, popular GPU models are often unavailable.
- Too few choices in GPU hardware, location, security level, latency, etc.
- Good GPUs are expensive, and projects can easily spend hundreds of thousands of dollars per month on training and inference.
io.net solves this problem by aggregating GPUs from underutilized resources such as independent data centers, crypto miners, and crypto projects such as Filecoin and Render, combined in the Decentralized Physical Infrastructure Network (DePIN), allowing Engineers can access more GPU computing power and pay lower fees.
io.net uses three products to involve users and computing power providers in this process to improve the user’s product experience:
IO Cloud
is a page for deploying and managing decentralized GPU clusters, where users can access a central hub for various functions, such as deploying clusters, browsing the GPU Marketplace, quickly monitoring operations, and getting started.
IO Cloud is also where users run AI/ML applications. It integrates seamlessly with IO-SDK to provide a comprehensive solution for scaling AI and Python applications.
IO Worker
Provides users with real-time insights into their computing, providing an operational and bird's-eye view of devices connected to the network, allowing them to monitor these devices and perform quick actions such as deleting and renaming devices. In addition to this, IO Worker can monitor computing activity, real-time data display, temperature and power consumption tracking, installation assistance, wallet management, security measures and profitability calculation related functions.
IO Explorer
Provides a window into the inner workings of the network, giving users comprehensive statistics and a bird's-eye view of all aspects of the GPU Cloud, with statistics such as network activity, vital statistics, data points and complete visibility of reward transactions.
Token economy
IO is the native token and protocol token of the io.net network. The main utility of tokens is in two aspects:
- IO will serve as the main payment method in the io.net ecosystem, such as paying for GPU deployment costs. And every model deployed on io.net must make a tiny IO transaction for inference.
- IO tokens are used to reward GPU contributors.
The network's goal at launch is to distribute 500,000 IO per month, with a two-year halving plan, and a maximum supply of IO of 22,300,000.
The IO token is expected to be launched in the first quarter of 2024.
2.BP-FLAC: AI-oriented public chain
BP-FLAC is the abbreviation of "blockchain-driven federated learning artificial intelligence", a generative AI infrastructure public chain. This project aims to use blockchain technology to aggregate and allocate GPU node computing power for training AI.
In order to protect data from different computing power contributors (miners), BP-FLAC uses the zk-SNARK algorithm to solve data privacy issues in AI training and provides token rewards for users who provide computing power support.
The BP-FLAC network consists of a hybrid blockchain model containing public and private chains.
Among them, the public chain adopts the PoC consensus mechanism, and miners contribute computing power (GPU resources) to help train the AI model and obtain corresponding rewards.
The private chain adopts the PoS consensus mechanism, and pledge nodes verify transactions and participate in protecting the network to obtain staking rewards. The more tokens you invest, the more rewards you will receive.
PoC (Proof of Certificate) means certificate proof. After miners provide GPU resources, they will obtain certificates as proof of their contribution. These certificates form the basis for the distribution of token rewards on BP-FLAC. The more computing power they contribute, the more rewards they receive.
Token economy
WOD Coin is the native token of BP-FLAC and is used to reward miners and nodes to facilitate transactions and protect the network through staking. The total supply is capped at 500 million and distributed as follows:
- 75% (375 million coins) will be distributed as mining (PoC) and staking (PoS) rewards over 80 years.
- Team members receive 5% (25 million coins), locked for two years, and linearly released within 48 months after two years (10 million for technology, 15 million for the founding team).
- 4% (20 million coins) will be used for the private placement round.
- 6% (30 million coins) will be used for marketing operations, locked for 6 months.
- 10% (50 million coins) will be used for ecological construction and locked for 6 months.
The official plan is to realize the incentive mapping from the test network to the mainnet in mid-2024, and reward early contributors when the mainnet is launched.
In terms of financing, the project completed US$10 million in financing from BP-FLAC on November 21, 2023, from Eureka Partners, Westlabs, and Mybitdata Ltd. , DecentraLabs, technology giants Amazon and NVIDIA, as well as Middle Eastern family office Alkhabeer Fund participated in the investment.
3.Gensyn: Distributed computing network for AI model training
Gensyn is a distributed computing network used to train AI models.
The calculations required for large-scale artificial intelligence are very complex, and with the continuous advancement of artificial intelligence (GPT3 to GPT4), more and more computing power is required. In the real world, due to various factors such as geopolitics, it is difficult to fully utilize global computing power and the cost has risen sharply. Some developers even have difficulty obtaining a single GPU, and they have to pay high service fees to the cloud server oligarchs.
In this context, Gensyn came into being.
Gensyn's concept is to connect the computing power of idle machine learning-capable computing devices (such as consumer GPUs, custom ASICs, and SoC devices capable of training neural networks) around the world through a global super cluster, thereby significantly Increase available computing power for machine learning.
This also brings good news to miners who have used ETH mining. With the transformation of ETH to PoS, the benefits obtained by many miners with powerful GPUs have plummeted. The emergence of the Gensyn protocol is expected to introduce the computing power of miners into AI.
The hourly cost of Gensyn's machine learning training work is about $0.40, which is far lower than the costs required by AWS ($2) and GCP ($2.50).
The Gensyn system mainly consists of four participants:
- Submitter: A system end user who provides tasks to be computed and pays for completed units of work.
- Solvers: The main working part of the system, performing model training and generating proofs for verification by the verifier.
- Verifier: Links the non-deterministic training process to the deterministic linear calculation, copies part of the solver proof, and compares distances to expected thresholds.
- Whistleblower: The last line of defense for result output, checking the work of validators and challenging them in hopes of winning the jackpot.
The four participants collaborated with each other to build Gensyn's verification system.
On June 11, 2023, Gensyn completed a $43 million Series A financing, led by a16z, with participation from CoinFund, Canonical Crypto, Protocol Labs, Jsquare, Eden Block and other angel investors.
4.MyShell: Easily create AI robots
MyShell is a platform for creating voice chatbots based on Web3 and AI. Users can select their favorite character at any time to start voice chat immediately and quickly improve their language skills by discussing topics of interest.
Code-free creation is at the heart of MyShell. MyShell hopes that through them, even users who are college students who have never learned programming can easily create their favorite AI robots.
How to do all this?
MyShell integrates two major model suites: LLM and Voice Cloning/TTS.
LLM, also known as large language model, is composed of an artificial neural network with many parameters. It uses self-supervised learning or semi-supervised learning to train a large amount of text, so that it can play any role, such as ChatGPT. MyShell uses LLM to make user-created AI characters more humane.
TTS (text-to-speech) and voice cloning technology are integral to voice communication with AI robots. Text-to-speech converts written text into spoken words, allowing users to access information audibly, while voice cloning replicates a person's unique voice characteristics, creating a personalized and authentic audio experience.
These two major suites enable MyShell to create a variety of humanized AI assistants. In May of this year, its total number of users exceeded 100,000.
In MyShell, income is generated mainly through the following four ways:
- Users pay monthly subscription fees in fiat or cryptocurrency;
- Creators pay to promote their AI-native apps to gain more exposure;
- Royalties from AI content/assets traded between users;
- AI developers purchase user-licensed data for model training and fine-tuning.
Token economy
$SHELL is the native token of the project, which can be used to pay the subscription fee of Shell NFT to obtain dialogue quota and Earn opportunities, as well as for community governance. The total supply is 1,000,000,000. 40% of it is used for community incentives, 20% is deposited in the ecosystem/treasury, 15% is used for private sales, and the team receives 17%.
Note: MyShell is currently in beta and the utility/distribution of tokens may change in the future.
On October 16, 2023, MyShell completed a $5.6 million seed round of financing at a valuation of $57 million, led by INCE Capital, with participation from Hashkey Capital, Folius Ventures, SevenX Ventures, OP Crypto and others.
5. Kaito AI: AI-powered encrypted information aggregation platform
Kaito is a Web3 search engine powered by AI that hopes to revolutionize cryptocurrency research and investment through AI.
For retail investors, the methods of obtaining information are extremely fragmented and unorganized. Web3's information index does not have a natively supported search engine and is easily limited by traditional searches represented by Google. To solve this pain point, Kaito appeared.
Kaito's search engine uses the Auto GPT framework and multiple ChatGPT backends to jointly build a proxy network that can handle various tasks, including search, information processing, data cleaning and annotation, aiming to provide higher quality Web3 information services . In addition to this, Kaito also leverages AI to optimize its search engine through ranking, topic mining, personalization, recommendations, speech-to-text transcription, and AI-generated content.
MetaSearch is Kaito's main product. Through MetaSearch, users can search all platforms in the encryption field with one click, including: Twitter, Discord, governance forums, Mirror, Medium, etc.
With MetaSearch, users can gain the following advantages:
- Search any cryptocurrency, topic or trend in seconds to get instant insights;
- Reduce time wasted on manual research with curated Twitter searches customized for Web3;
- Early detection of narratives through social signals;
- Through AI, any technical documents and papers of users can be answered instantly.
Currently, the basic product is priced at $129, which works out to $99 per month on an annual basis.
On June 22, 2023, Kaito completed a US$5.5 million Series A financing with a valuation of US$87.5 million, led by Superscrypt and Spartan.
Token economy
There is currently no specific token information for this project. You can pay close attention to its official website for the latest trends.
6.Aethir: A cloud gaming platform that solves delays and piracy
Aethir is a real-time rendering network built specifically for games and artificial intelligence. Through the cloud services and rendering technology it provides, it ensures that gamers can get an ultra-low latency experience anywhere and on any device.
The Aethir cloud gaming platform is its primary venue for providing this service. Behind the Aethir cloud gaming platform is a large and stable distributed computing node network, allowing games to run directly in the cloud. As a result, players can enjoy a high-quality, seamless gaming experience on any device without downloading.
For game developers, Aethir can be used to quickly deploy and launch their games online without going through complex development and adaptation processes.
Within the Aethir platform, "games" are viewed as a collection. Within this collection, multiple "game versions" can be created. For each version in the "Games" collection, users can submit, modify or delete game adaptation services. Therefore, Aethir's cloud platform not only allows developers to seamlessly implement updates, but also ensures the synchronization of game versions for all users.
Aethir’s cloud platform also solves the problem of widespread piracy in the gaming industry.
In traditional games, pirated games have always been a thorn in the side of the gaming industry, but with Aethir this problem is easily solved. This is because the game is executed on a cloud server rather than on the user's device, and the user has no access to any game files, so it eliminates almost all forms of traditional game piracy.
As Arthur Hayes explains, Aethir is meaningfully expanding the addressable market for games and other streaming content in the world's fastest-growing (but hardware-constrained) markets. As Aethir scales, everyone benefits – gamers, publishers, and globally distributed render node operators alike.
Aethir's audience is not only players within Web3, but all players from all walks of life around the world.
On July 27, 2023, Aethir completed a $9 million Pre-A round of financing, valued at $150 million, led by investors such as Sanctor Capital, Hashkey, Merit Circle and CitizenX.
Token economy
There is currently no specific token information for this project. You can pay close attention to its official website for the latest trends.
7.Ritual: Integrating AI into the execution layer of smart contracts
Ritual is an AI-based sovereign execution layer. Their vision is to enable developers to build fully transparent DeFi, self-improving blockchains, autonomous agents, generated content, and more through Ritual.
Infernet is the core suite of Ritual. It is essentially a lightweight library that introduces calculations to the chain. Its main application is to serve ML inference workloads. Users can build and host ML models, deploy them to Infernet nodes, and create subscriptions to consume inference outputs and optional concise proofs of execution from on-chain smart contracts.
Infernet SDK can be used for integration between Web3 and AI functions, such as: data preprocessing and training, model support (scikit-learn, HuggingFace), verification (zero-knowledge proof, Optimistic fraud) proof, various open source zk libraries), data Escrow and provenance, etc.
Through the flexible application of Infernet SDK, developers can integrate artificial intelligence models into users' own protocols, applications or smart contracts. Enables developers to fine-tune, monetize, and infer models using cryptographic schemes.
Token economy
There is currently no specific token information for this project. You can pay close attention to its official website for the latest trends.
On November 8, 2023, Ritual completed a $25 million Series A financing led by Archetype.
