Author: DODO Research
The combination of AI and blockchain is currently a hot topic in the industry, with developers actively exploring the possibilities of this integration. Blockchain technology is being seen as an ideal choice to address various issues such as the inability to effectively utilize AI services and computing resources. Many projects have already carved out their own path in this field.
Today, Dr. DODO will take you through some projects that are performing well in the AI and computing power market track.
AI + Blockchain
Projects focusing on AI in the blockchain field can be primarily divided into three tracks:
1. Computing power sharing: Blockchain technology can build a distributed cloud computing platform, enabling the sharing and efficient utilization of computing resources. Smart contracts can lease out idle computing resources to those in need, thereby improving resource utilization and reducing costs. Representative projects include io.net and Aethir.
2. AI data security and verifiable computing: Both AI and blockchain technologies deal with large amounts of data. Blockchain storage networks offer secure data storage and transmission methods, while AI analyzes this data to generate valuable information. Integrating these two can protect user privacy while providing reliable data sources for AI. Currently, projects like Arweave represent this direction.
3. Decentralized AI: Deploying AI models on blockchain networks can provide decentralized AI services, enhancing system reliability and stability while reducing the risk of single point failures. Projects like Bittensor represent this direction.
io.net
Recently launched on Binance, io.net is currently a standout project in this track. io.net is a decentralized GPU network designed to provide significant computing power for machine learning applications. Their vision is to unlock fair computational access by assembling over a million GPUs from independent data centers, encrypted miners, and projects like Filecoin, making computing more scalable, accessible, and efficient.
io.net offers a completely different cloud computing approach, utilizing a distributed and decentralized model to provide users with more control and flexibility over computing power. Their services are permissionless and cost-effective. According to io.net, their computing power is 90% lower than centralized providers like Amazon AWS, making them a standout decentralized provider.
Aethir
Aethir offers a disruptive yet highly feasible solution to effectively utilize global computing power resources. Their network aggregates and intelligently redistributes new and idle GPUs from enterprises, data centers, cryptocurrency mining operations, and consumers. The market opportunities for better reallocating GPU capacity are vast, and Aethir aims to increase global GPU computing availability by over 10 times.
A key feature of Aethir is its focus on reusing existing idle resources rather than requiring node participants to purchase new hardware. Typically, the underutilization of GPU capacity in a device ranges from 50% to 75%, indicating a significant amount of computing resources that can be tokenized. Aethir aims to leverage these abundant idle resources by targeting small to medium data centers and enterprises.
Aethir’s token has been listed on exchanges like OKX and Bybit after raising $9 million in Pre-A funding from prominent institutions like Sanctor Capital and Hashkey.
Arweave
AO is a distributed, decentralized, participant-oriented computing system based on Arweave. The core goal of AO is to provide a trustless, collaborative computing service with no practical scale limits, offering a new paradigm for applications integrated with blockchain. Unlike other high-performance blockchains, AO supports storing large amounts of data, such as AI models. Unlike Ethereum, AO allows an arbitrary number of parallel processes to run within a computing unit, coordinating through open message passing without relying on centralized memory space.
Arweave’s launch of AO signifies a move from the decentralized storage track to a broader decentralized cloud services arena. Their permanent on-chain storage aims to become a permanent host for cloud computing, focusing on large-scale verifiable computing.
Arweave recently announced the token economy between $AR and $AO. According to official statements, $AO is a 100% fairly distributed token with no presale or allocation. The total supply of $AO is 21 million, with a halving cycle of 4 years, distributing every 5 minutes, with a monthly distribution of 1.425% of the remaining supply.
Approximately 36% of $AO tokens are distributed to $AR holders every 5 minutes, incentivizing the security of AO’s base layer – Arweave.
The remaining approximately 64% of $AO tokens are allocated to bridging users, providing external revenue and incentivizing asset introduction to AO.
Bittensor
Training AI models requires significant data and computing power, but the high costs have led to these resources being monopolized by large enterprises and research institutions. This centralization restricts the use and collaboration of AI models, hindering the development of the AI ecosystem. Bittensor (TAO) aims to establish the world’s first blockchain neural network to facilitate the exchange of machine learning capabilities and predictions among network participants.
Bittensor hopes to promote the sharing and collaboration of machine learning models and services in a peer-to-peer manner. While technically challenging, TAO is still a distance away from practical application.
Author’s Perspective
These AI + blockchain projects are likely to change the future landscape of computing resource allocation. Decentralized ownership, collaborative cross-cluster decentralized regional deployment, will pave the way for a new wave of economic and technological advancement. These projects have ambitious goals, aiming to reshape the landscape of future cloud computing and AI applications, creating a more interconnected, efficient, and innovation-driven global cloud economy. With countries actively promoting productivity transformation, these development directions are worth exploring in depth. However, due to the need for stronger technical support and more financial backing in this field, the barrier to entry for project teams is not low. Currently, they are still in the trial phase, and whether they can be implemented as practical infrastructure that people will actually use remains to be seen.
