GPU Powered Inference Now Available on Testnet
Artificial intelligence has rapidly become one of the most important drivers of innovation, and blockchains are increasingly used to add transparency, permanence, and verifiability to AI-related workflows. Up until now, however, most “on-chain AI” solutions have been limited to recording interaction data, while the actual model hosting and execution happens off-chain.
The goal of bringing AI inference on chain drove the original design work on the AI Inference Extension, which was explored throughout 2025. The first version demonstrated that lightweight models could be executed on standard Chromia nodes using CPU resources. While this confirmed the feasibility of the approach, it also made clear that running larger, more capable models would require greater computational capacity.
Enter GPU Enabled Nodes
Graphical Processing Units (GPUs) are the workhorses of modern AI. Their architecture is essential for performing the large-scale matrix operations that power today’s large language models (LLMs). Bringing GPUs into the Chromia node ecosystem opens the door for running far more capable models directly within the blockchain’s execution environment.
In the past months, we have developed the infrastructure needed to make this a reality. Specialized Chromia nodes equipped with GPUs are now capable of performing these heavier computations, enabling real model execution inside the blockchain’s execution layer rather than relying on external systems.
Introducing the Updated AI Inference Extension
Today, we are presenting a public demonstration of LLM inference executed on-chain. The demo can be accessed here and provides a concrete example of how AI workloads can be executed directly within Chromia’s distributed architecture.
The current implementation runs the Qwen2.5-1.5B-Instruct model on GPU enabled nodes operating on Chromia Testnet.
This demonstration should be understood as a technical proof of concept. It shows that AI inference can be executed on-chain, with model inputs and outputs recorded as transaction data that can be inspected and verified directly on the network.
The model used in this demo can respond clearly and coherently to prompts, showing that meaningful AI inference can run directly within Chromia’s architecture. At the same time, it does not have access to up to date information, and its responses are limited to its original training data and the context of each interaction. This demo is not a polished product or a ChatGPT equivalent, but rather an illustration of the potential of on-chain inference and a concrete starting point for developers exploring what they can build with it.
When Mainnet?
These testnet nodes use the same Postchain based execution framework as mainnet, meaning that future deployment is largely a matter of configuration and rollout. While the demo represents meaningful progress, deploying GPU enabled nodes to mainnet requires careful consideration in a few key areas.
1. Economics
GPU hardware is more expensive to set up and maintain than standard validator configurations. These nodes also consume more power, leading to higher operational costs. These factors need to be weighed when determining the size and scope of a mainnet GPU cluster.
2. Demand & Market Readiness
There are several use cases that benefit from the tamper resistance, transparency, and verifiability that decentralized LLMs can provide, including public records, government services, regulated industries, and public-interest applications.
At the same time, awareness of decentralized AI execution options is still developing, largely because the technologies are new. Adoption will take time, particularly in sectors that depend on clearer regulatory guidance, strong risk management practices, and a high degree of public trust.
3. Technical, Product and Market Coordination
Chromia is planning new data and AI-focused initiatives for next year. Introducing GPU-enabled nodes into the ecosystem requires coordination across these efforts, including how these nodes can best integrate with and support other upcoming products and capabilities.
The Road Ahead
AI Inference v2 and GPU-enabled nodes form the foundation for more capable decentralized applications. They allow AI models to run directly within Chromia’s execution environment, expanding what is possible on-chain. Development will continue to refine these capabilities and integrate this infrastructure with other emerging products and initiatives.
About Chromia
Chromia is a Layer-1 relational blockchain platform that uses a modular framework to empower users and developers with dedicated dapp chains, customizable fee structures, and enhanced digital assets. By fundamentally changing how information is structured on the blockchain, Chromia provides natively queryable data indexed in real-time, challenging the status quo to deliver innovations that will streamline the end-user experience and facilitate new Web3 business models.
Website | X | Telegram | Instagram | Youtube | Discord | Reddit | LinkedIn | Facebook | BlueSky