Developers working on the Internet Computer are experimenting with running language models entirely on-chain, with early results showing the approach is technically possible even if the experience remains limited compared with mainstream AI systems.
Bitcorn Labs is among the groups pushing the idea forward. The team has been running advanced models directly inside canisters, the Internet Computer’s smart contract containers. That setup means the model, its data and the inference process all operate within the blockchain environment.
Current deployments still face practical limits. Canisters have maximum size restrictions, which caps how large a model can be. Running workloads on the network’s typical 13 node subnets also carries a high cost, making large scale use difficult at this stage.
Even so, developers involved say the experiments show the concept works. Improvements to the network, including AI inference subnets or configurations with fewer nodes, could change the economics and performance over time.
For now, the models resemble the small local AI systems many developers were experimenting with a year or so ago. They can produce responses and complete basic tasks, though conversations tend to feel slow and the reasoning ability remains modest.
Supporters of the approach argue that human facing chat applications are not the only use case. Autonomous software agents operating on-chain may not require the same speed or capability as consumer AI tools. Agents can tolerate slower responses, and tasks may only require a model that performs adequately rather than one designed for complex reasoning.
The work by Bitcorn Labs builds on earlier efforts by the developer group Onicai, which explored ways to run language models directly on the Internet Computer. Researchers in the community say combining on-chain infrastructure with AI agents could open new forms of automated applications that interact with decentralised finance tools, smart contracts and other blockchain services.
One example circulating in the community comes from a project called Waifu.ic0.ai. The team said it has deployed an on-chain language model connected to an agent using a Dfinity LLM canister.
According to the developers, the system links with contributors and tools across the Internet Computer ecosystem, including developer Snassy ICP and the trading platform Sneedex. Users can manage a portfolio connected to the “waifu” system, operate a personal bot and generate small pieces of code.
The developers say the system is evolving quickly and describe the current release as an early version.
Advocates of fully on-chain AI argue the approach could bring transparency to automated decision making, since both the models and their actions would run within the blockchain environment. Critics say the costs and performance limits remain serious obstacles when compared with conventional cloud based AI systems.
Experiments like those from Bitcorn Labs and Waifu.ic0.ai suggest developers are willing to test the idea despite those hurdles. Whether the approach becomes practical at scale will depend on improvements in infrastructure, model efficiency and the economics of running AI workloads directly on blockchain networks.
Dear Reader,
Ledger Life is an independent platform dedicated to covering the Internet Computer (ICP) ecosystem and beyond. We focus on real stories, builder updates, project launches, and the quiet innovations that often get missed.
We’re not backed by sponsors. We rely on readers like you.
If you find value in what we publish—whether it’s deep dives into dApps, explainers on decentralised tech, or just keeping track of what’s moving in Web3—please consider making a donation. It helps us cover costs, stay consistent, and remain truly independent.
Your support goes a long way.
🧠 ICP Principal: ins6i-d53ug-zxmgh-qvum3-r3pvl-ufcvu-bdyon-ovzdy-d26k3-lgq2v-3qe
🧾 ICP Address: f8deb966878f8b83204b251d5d799e0345ea72b8e62e8cf9da8d8830e1b3b05f
Every contribution helps keep the lights on, the stories flowing, and the crypto clutter out.
Thank you for reading, sharing, and being part of this experiment in decentralised media.
—Team Ledger Life
Community Discussion