AITECH Cloud Network

What is AITECH Cloud Network?

AITECH Cloud Network is an infrastructure network designed for AI development and enterprise automation; its core components include on-demand GPU computing, AI agent orchestration, and blockchain-based access and settlement mechanisms.

Previously, this project used the Solidus AI Tech brand; by 2026 it will complete the transition to AITECH Cloud Network, with the original AITECH tokens being replaced by ACN tokens on the Ethereum network.

The core components of the platform

1. Compute Layer GPU computing power market

The Compute Layer offers GPU instances for model training, inference, rendering, and scientific computing. Users can select the hardware and duration of use based on their actual needs, without having to purchase physical servers or enter into long-term cloud subscriptions.

• Hardware options such as H100, H200, B100, B200, B300, L40S, and the RTX 6000 series are available.
• It supports creating GPU instances on demand, as well as scaling up or shutting down resources based on project needs.
• Designed for tasks such as large model training, inference services, CGI rendering, and research simulations.
• Data center operators and GPU suppliers can apply to list their unused computing power, specifying the available hours and prices.

2. Agent Forge AI agent orchestration

Agent Forge is a visualization-based, natural-language-driven AI workflow platform. Users can connect models, tools, triggers, and output nodes on a canvas, without having to write all the integration code from scratch.

• It supports model sources such as OpenAI, Anthropic, Google, Groq, Cerebras, and Ollama.
• Workflows can be executed through the chat interface, REST API, Webhook, scheduled tasks, or external events.
• It offers integrations for email, collaboration, databases, search, documents, e-commerce, and development tools.
• It supports logging, monitoring, error handling, and multi-user collaboration.
• Workflows can be deployed as APIs or registered as discoverable, chargeable proxies.

3. AI Marketplace

The AI Marketplace serves to bring together foundational models, middleware, plugins, AI applications, datasets, and proxy workflows. Developers can publish solutions, while users can select models or applications and deploy them using available computing resources.

The market may contain open-source models and projects contributed by the community, but the platform itself does not guarantee the security, accuracy, license compatibility, or regulatory compliance of each project.

4. ACN tokens and network mechanism

ACN is the ecosystem token currently used in networks; it can be employed for partial top-ups, access to services, staking, agent deployment, and on-chain settlements. Agent Forge also plans to integrate ERC-8004 for agent registration along with the x402 payment mechanism.

Tokens are volatile crypto-assets; their value is not fixed, nor does it correspond to a specific return on investment. Users should only operate through approved official channels, and they must themselves determine the legal requirements related to tokens, staking, and blockchain services in their respective regions.

What can GPU computing power be used for?

• Train, fine-tune, and evaluate machine learning or large language models.
• Deploy text, image, voice, and multimodal reasoning services.
• Execute batch data processing, research simulations, and high-performance computing tasks.
• It is used for 3D rendering, video processing, and the creation of generative content.
• Provides on-demand model inference capabilities for the Agent Forge workflow.

In what scenarios is Agent Forge suitable?

• Customer support: Connects to a knowledge base, ticketing system, email, and chat tools to automatically handle common requests.
• Sales and marketing: studying customers, creating content, updating CRM, and initiating subsequent communications.
• Data processing: extracting document information, cleaning tables, performing searches, and generating reports.
• E-commerce operations: Linking products, orders, customer service, and messaging channels to create automated processes.
• Development and operations: Coordinate multi-step tasks using Webhooks, databases, and monitoring tools.

How to rent GPU computing power?

1. Create an account and complete the identity or corporate verification required by the platform.
2. Confirm the required VRAM, GPU model, number of nodes, runtime, and network storage for the task.
3. Top up with fiat currency or cryptocurrency, and verify the currency type, blockchain network, method of settlement, and fees.
4. Select the GPU resources and image environment to view the hourly cost per card and the estimated total cost.
5. After creating the instance, configure the access keys, network rules, storage, and runtime environment.
6. First, conduct short-duration benchmark tests to verify performance, drivers, frameworks, and data transfer speeds.
7. Launch the task officially and continuously monitor utilization, logs, budget, and any abnormal conditions.
8. After the task is completed, download the data, delete the keys, and shut down the instance to prevent unnecessary charging for idle resources.

How to use Agent Forge?

1. Choose a blank workflow or one of the pre-set templates for marketing, customer service, sales, HR, etc.
2. Describe the goal in natural language, or add model, agent, and tool blocks to the visualization canvas.
3. Connect to email accounts, documents, databases, CRM systems, or search services, applying the principle of least privilege.
4. Set the input, branching conditions, failure handling, output format, and manual approval steps.
5. Run the workflow using test data, and check the logs, execution time, and costs for each step.
6. After confirming that sensitive actions are subject to permission restrictions, deploy them via chat, Webhook, API, or a timer.
7. Continuously monitor the results, and set alerts for model errors, changes in third-party interfaces, and abnormal costs.

Public GPU price

As of August 25, 2026, the Compute Layer official website lists the following starting prices for single-card instances. These prices can vary depending on configuration, region, supply and demand, storage capacity, data transfer volume, and cluster services; the final price will be indicated on the page for creating an instance or requesting a quote.

The website lists the B100, B200, B300, and RTX 6000 series, but no fixed public prices are displayed for all of these models upon verification. Bare-metal clusters are configured according to specific requirements, and quotes are obtained from the sales team.

Agent Forge price plans

Agent Forge uses a combination of subscription fees and usage-based charges. The table below shows the basic options available as of the verification date; additional models, third-party APIs, transactions on the blockchain, and GPU computing power may incur separate costs.

\"Unlimited usage\" is still generally subject to rules regarding fair use, model availability, concurrency, and platform security. Before making a purchase, it is necessary to understand the billing procedures in case the quota for inference is exhausted, as well as the options related to monthly resets, refunds, renewals, and ACN benefits.

Key points for cost estimation

• The hourly cost of a GPU represents only the basic computing cost; storage, data transfer, and the time required to prepare the images also need to be taken into account.
• For multi-card tasks, the cost is calculated by multiplying the number of cards by the actual runtime, with an additional amount set aside for attempts to retry in case of failure.
• The amount of inference credits included in the Agent Forge subscription does not mean that external models and third-party tools are all available free of charge.
• When using the built-in API key, the fees may be charged directly by the corresponding model or data service.
• Recharging or making payments on the chain can also give rise to network fees and fluctuations in asset prices.

Token migration and fraud prevention reminders

The original AITECH tokens have been transferred to ACN on Ethereum at a 1:1 ratio; the official migration portal indicates that the deadline is May 13, 2026. Old tutorials, old contract addresses, and information related to the old blockchain should no longer be used as a basis for current operations.

• Do not connect your wallet using private messages, search ads, or group chat links.
• First, access the service through the official website, and then verify the network and contract details.
• Run a small-scale test before sending, and verify that the receiving address matches the selected chain.
• Never reveal your mnemonic phrase or private key to customer service or so-called administrators.
• For special handling after the migration deadline, contact official support directly; do not rely on third-party services.

Privacy, KYC, and security

Some services require completion of KYC or KYB; identity documents, addresses, and biometric checks may be handled through third-party verification services. The platform also processes account, contact, transaction, device, and usage data.

Sensitive AI workloads also involve training data, model weights, prompts, and outputs. Before uploading such data, enterprises should determine the classification of the data, its storage location, retention period, encryption methods, access logs, as well as backup and deletion procedures.

• Use a separate key for GPU instances, and restrict the source addresses of the management ports.
• Do not include the wallet private key, model keys, or production credentials in the image.
• Manual approval is required for actions such as emails, payments, transactions, and deletions within the workflow.
• Check the separate privacy policies of third-party models, data sources, and integration tools.
• After stopping the service, delete instances, volumes, snapshots, tokens, and any authorizations that are no longer needed.

Open-source status and GitHub details

The commercial platform of AITECH Cloud Network, its computing power market, and the Agent Forge hosting services are not part of an open-source project as a whole. The official GitHub repository contains example repositories such as a Starter Kit for demonstrating basic model calls, but this does not mean that the core platform code is made available openly.

The AI Marketplace may also offer open-source models or community projects. Users should examine individually the maintenance status of the repositories, the licenses, the terms related to the model weights, and the security of the dependencies; they should not assume that just because there is open-source content available in the marketplace, the platform itself is open-source.

Product advantages

• Place GPU computing and AI agent workflows within the same ecosystem.
• The public GPU options have an hourly starting price, which makes it easy to estimate costs initially.
• Agent Forge supports natural language, visual orchestration, and various execution methods.
• It covers multi-model, data, communication, search, and business system integration.
• It enables providers of computing power to enter the market, thereby increasing the utilization rate of idle GPUs.

Usage restrictions and risks

• The actual availability, performance, and geographic coverage of decentralized resources can vary depending on supply and demand.
• The costs related to GPUs, models, third-party APIs, and on-chain fees are dispersed, so the actual total cost needs to be calculated separately.
• AI agents may make erroneous decisions or carry out unexpected actions, and essential human supervision cannot be eliminated.
• The community models and applications are provided as they are, and no guarantee is given that they are suitable for regulated scenarios such as healthcare, finance, and recruitment.
• Cryptocurrencies are subject to risks related to price fluctuations, smart contracts, regulation, wallets, and irreversible transactions.
• Certain regions or services may not pass the KYC, sanctions, or acceptable use policy checks.

Frequently Asked Questions

Is AITECH Cloud Network an ordinary cloud server?

It focuses more on GPUs, AI training and inference, agent workflows, and the distributed computing market. If low-cost CPU servers or traditional website hosting are what’s needed, general-purpose cloud services should also be considered.

Is it possible to rent a GPU without using ACN tokens?

The official website states that it is possible to set limits for encryption or fiat currencies, but the specific options available may vary depending on the region, payment methods, and services. It is necessary to check what options are available in the account before creating an instance.

Does Agent Forge require coding?

Basic workflows can be created using natural language and visual interfaces. However, complex custom APIs, data transformations, permission management, and production deployment still require expertise in development and operations.

Is ACN a worthwhile investment?

This page only explains the uses of tokens on the platform; it does not provide any investment advice. Token prices and relevant regulations can change rapidly, so users should conduct their own research and assume only the risks they are willing to take.

Is the platform open source?

It is not fully open source. The example code made available by the developers and the open-source models available on the market are merely components of this ecosystem; the core hosting platform remains a commercial service.