This site (120bTC.coM): Ankr, a blockchain infrastructure development company, announced the launch of the Neura test network, a blockchain network specially designed for allocating GPU resources for artificial intelligence and machine learning. The ANKR token will be adopted as Neura’s native cryptocurrency and the Neura mainnet is expected to be launched later this year.
What problem does Neura want to solve
Neura’s goal is to enable developers to build decentralized products that combine AI, cloud computing, and blockchain technology. Specifically, Neura will start by helping new startups that need machine learning computing resources to raise funds, obtain GPU resources, and store data.
Computing resources are in short supply
GPUs for model calculations have been in short supply for a long time due to changes in industry trends in recent years, and it is a seller's market for cloud computing manufacturers. This leads to increased costs for development teams to train AI models, and there are no other customization options.
The model lacks a revenue model
In addition to technical obstacles related to infrastructure scalability, AI and cloud computing projects also need to think about how to build a business model with sustainable revenue.
Traditional Web2 platforms and processes often fail to adequately support strategies for monetizing training materials, which is critical to the survival and growth of AI projects. If the model is open source, it will not be able to generate revenue, but if it is not open source, it will face more stringent regulatory inspections and less easy access to training data. Becomes the current economic dilemma of machine learning model development.
Introduction to Neura design concepts
Neura infrastructure
Neura is a blockchain built based on the Cosmos SDK. The network is compatible with EVM and is used by GPU providers Registration, registration of users who need to use computing services, ANKR’s token economic distribution model based on computing results, etc.
In addition, by integrating off-chain storage solutions such as Celestia, EigenLayer, IPFS and Arweave to store AI model data and computing results of GPU providers, that is, most AI operations are performed off-chain (operations are performed on The GPU provider's server and data storage are in infrastructure such as Celestia), and the Neura chain only handles the incentive model.
Neura design architecture
Neura operation process
In the Neura platform, the operation process is as follows:
When a task provider (such as a team that needs computing resources) proposes a task (such as the computing requirements of an AI model), the task provider can choose the most suitable GPU provider for itself and obtain its selected GPU provider. 's public key.
The task provider uses the public key of the GPU provider to encrypt the task content, and then submits the encrypted task content as a transaction to off-chain facility storage, such as Arweave and EigenDA.
The task owner creates a new transaction on the Neura network through RPC, which contains the ID of the previous transaction, the ID of the selected GPU Provider, the duration the task should be maintained, and the Specific requirements for GPU, RAM, CPU.
After the transaction content is confirmed by the verifier on the Neura network card, the corresponding smart contract on Neura will be updated.
The selected GPU provider detects the relevant transaction through the monitoring contract, verifies it, retrieves the task from the off-chain storage facility and executes it.
After the GPU provider completes the execution of the task, the results will be uploaded to the off-chain storage facility as transaction content. (The results can be unencrypted, depending on the privacy and confidentiality requirements of the task provider)
The GPU provider will establish another transaction to Neura to update the task status and provide the results A reference ID that is given to validators on the Neura network for confirmation.
Neura operation flow chart
Neura Token Economics
Use ANKR token
Neura will continue to use the ANKR token as its native token. When the Neura mainnet is online in the future, the Ankr team will provide a cross-chain bridge to allow holders to transfer ANKR tokens to the Neura blockchain.
ANKR tokens will play the role of paying for GPU resources on the Neura network, participating in IMO, and serving as a reward for allocating GPU suppliers.
Profit sharing model mechanism
In the Neura network, task owners can issue ERC‑7641 tokens to share the income from creating AI models. These tokens allow holders to obtain Generate a portion of the revenue to ensure a fair reward distribution system for contributors and investors.
In addition, ERC-404 will be introduced to share the ownership of the model with Shengta participants as NFT, promoting task providers to conduct the initial token issuance of the project through crowdfunding.
The team stated that these mechanisms can leverage Neura blockchain transparency to cultivate bilateral trust and produce a sustainable computing ecosystem.
AI Cloud Computing Competitors
In addition to Neura introduced this time, io.net, Akash Network and even Near on Solana have joined the competition. Decentralized computing infrastructure is flourishing, but competition is becoming increasingly fierce. Now that Ankr, which has narrow resources, has joined the fray, it is bound to create more competition in the industry.
With traditional Web2 services not formally familiar with the development of the encryption ecosystem, the existing computing needs of these projects are limited. How to compete with traditional cloud computing platform players including AWS, Google Cloud and other players, and let outsiders compete? Teams with computing needs see it? It will be a bigger challenge that decentralized GPU computing platforms such as Neura will face.
The above is the detailed content of Ankr’s upcoming decentralized AI computing network Neura! Continue to use ANKR token. For more information, please follow other related articles on the PHP Chinese website!