Giza

Giza

Created using Figma
Giza is a Swiss‑based Web3 infrastructure startup (founded in 2022) focused on trust‑minimized integration of machine‑learning (ML) into decentralized finance and smart contracts. With under ten employees and backed by investors like CoinFund, StarkWare, Arrington, MH Ventures, and Base Ecosystem Fund, Giza aims to democratize on‑chain ML by providing a full-stack platform built atop StarkNet’s ZK‑rollup architecture.
To be announced
Additional Details
Categories
Platform

About Giza

Giza is a Switzerland-based Web3 infrastructure startup founded in 2022, aiming to integrate machine learning (ML) into decentralized systems through trust-minimized architecture. Built on StarkNet’s ZK-rollup technology, Giza enables developers to deploy and run ML models directly on-chain, eliminating the need for centralized cloud infrastructure. The platform combines cryptographic security, scalability, and composability to bring AI-native applications into the blockchain ecosystem.

At the heart of Giza's stack is the ability to deploy ONNX-compatible ML models to the StarkNet network. These models can execute inference on-chain, ensuring high uptime, automatic scaling, and transparent operation. This marks a major shift from off-chain computation toward verifiable, blockchain-native intelligence.

Giza provides a command-line interface (CLI) that allows developers to package their models into “AI Actions.” These actions are verifiable units of computation that can interact directly with smart contracts. On top of this, Giza introduces “Agents”—modular, stateful automation layers that give models memory, monitoring, and reflective capabilities. Agents can respond to events, update their internal logic, and execute decisions autonomously on-chain.

The platform supports a wide range of ML frameworks such as TensorFlow, PyTorch, and Scikit-learn by using the ONNX standard. It enables applications across DeFi, gaming, DAOs, and oracles, where intelligent behavior and real-time decision-making are crucial.

Giza’s infrastructure is open-source and developer-friendly. The team has released essential tools like the CLI, agents framework, and the Orion runtime to foster community collaboration. The company also launched a native token, $GIZA, to reward early users, contributors, and developers building in the ecosystem.

Features

Simplified ML Ops
Secure deployment of serverless on-chain models with no infrastructure maintenance.

On-chain Inference
Native on-chain inference enables trustless, verifiable, and transparent model usage.

Ownership & Monetization
Transferable ML models grant model owners rights to fees per usage.

Giza Roadmap

  • Q2 2023

  • Foundation
    Development of foundational tools and architecture in close collaboration with partner protocols.

    Public Release
    Permissionless machine learning model deployment and on-chain inference.
  • Q4 2024

  • Open AI Marketplace
    A marketplace that provides an alternative path for the collective development of AI.

Participation in events

Giza Team

Verified 0%

Attention. There is a risk that unverified members are not actually members of the team

Renç Korzay
CEO: Former Head of Product at Johnson & Johnson a...
unverified
Fran Algaba
CTO: Former Head of ML Engineering at BBVA and Adi...
unverified
Cem F Dagdelen
CPO: Previously founded Curve Labs. Extensive mech...
unverified
N/A
5.0 20
WEB3 Profile Vision Activity Potential Product Team
  • Due to potential time differences in information updates, please verify the accuracy of each ICO project through its official website or other official communication channels.
  • This information is not intended as a recommendation or suggestion for ICO investment. Please conduct thorough research on the relevant information and make your own informed decision regarding ICO participation.
  • If you identify any issues or errors in this content, or if you wish to submit your own ICO project for listing, please contact us via email at info@icoholder.com.
Please read the disclaimer and risk warning. Show disclaimer and risk warning.