Can Live Data Redefine Intelligence?

At Token2049 Singapore’s Real World AI Summit, IoTeX and a coalition of industry partners unveiled the Real-World AI Foundry.

The global initiative aims to define shared standards for artificial intelligence grounded in live, verifiable data.

IoTeX Pushes Boundaries of AI and Blockchain Integration

Sponsored

Sponsored

The Foundry seeks to tackle limitations typical to AI, including reliance on static, historical datasets that power prediction but lack adaptability to real-world environments.

Real-world models (RWMs) are central to the project. They operate as multimodal AI systems designed to integrate data from machines, sensors, and human interaction.

Unlike large language models (LLMs) or traditional machine learning trained on archives, RWMs aim to perceive live events, learn cause-and-effect, and respond dynamically.

“IoTeX and its partners are attempting to move AI from abstract prediction to grounded action. The Real-World AI Foundry brings together enterprises, researchers, AI innovators, and infrastructure providers to co-architect an open foundation for intelligence that is live, verifiable, and aligned with human values,” said Raullen Chai, Co-Founder and CEO of IoTeX, in a statement shared with BeInCrypto.

The initiative has attracted a broad alliance. Partners range from Vodafone and the Blockchain Association to crypto-native projects, including Theta Network, Filecoin (FIL), Aethir, Hashkey Chain, Drone Layer, and Taiko.

Collectively, they have pledged to build under three guiding principles:

• Grounded — RWMs must be based on verified, real-time data.
• Open — Frameworks should be interoperable and open-source.
• Human — Governance must embed accountability and alignment with human values.

Why the Timing Matters

Sponsored

Sponsored

The launch comes as autonomous systems begin to influence sectors such as healthcare, energy, and mobility. Reliance on outdated or misaligned models poses increasing risks in these sectors.

Traditional AI often struggles with context, novelty, and accountability, which are shortcomings that RWMs could address by training on live, trusted data.

Still, challenges remain. Achieving consensus across industries, safeguarding data ownership, and addressing liability in autonomous decision-making will likely spark debate. Critics may question whether a decentralized alliance can deliver coordination at scale.

Meanwhile, it is worth mentioning that the Foundry will not operate as a centralized entity. Instead, participants will contribute specialized inputs, ranging from infrastructure and data streams to research expertise and adoption pathways.

At the same time, they will collectively shape governance, quality standards, and interoperability requirements.

If successful, the initiative could establish Real-World AI as a new category alongside LLMs and generative AI. Its emphasis on multi-stakeholder governance contrasts sharply with Big Tech’s closed, proprietary approaches.

The Real-World AI Foundry is still in its early stages, and many technical and governance frameworks have yet to be finalized.

As of this writing, IOTEX, the network’s powering token, was trading for $0.0231, down by 1.12% in the last 24 hours.