Trusted, Scalable AGI on Verifiable Open Hardware

chevron-icon
RFP Proposals
Top
chevron-icon
project-presentation-img
user-profile-img
Vinay Devabhakthuni
Project Owner

Trusted, Scalable AGI on Verifiable Open Hardware

Expert Rating

n/a

Overview

We propose a novel AGI hardware stack combining RISC-V microarchitectures, symbolic reasoning runtimes and zero-knowledge proof systems (RISC Zero). This enables decentralized, verifiable AGI agents to run on edge devices like phones, drones or sensors. Our stack supports MeTTa logic, outputs cryptographic proofs of decision-making and includes an SDK for SingularityNET integration. Optional neuromorphic hooks enable hybrid reasoning. All components are open-source, scalable and optimized for low-power, trustless deployment.

RFP Guidelines

Explore novel hardware architectures and computing paradigms for AGI

Complete & Awarded
  • Type SingularityNET RFP
  • Total RFP Funding $80,000 USD
  • Proposals 9
  • Awarded Projects 1
author-img
SingularityNET
Apr. 14, 2025

The purpose of this RFP is to identify, assess, and experiment with novel computing paradigms that could enhance AGI system performance and efficiency. By focusing on alternative architectures, this research aims to overcome computational bottlenecks in recursive reasoning, probabilistic inference, attention allocation, and large-scale knowledge representation. Bids are expected to range from $40,000 - $80,000.

Proposal Description

Our Team

Our team brings interdisciplinary expertise across:

  • Systems Architect

  • RISC-V Engineer

  • ZKP Engineer

  • Runtime Developer

  • DevOps Engineer

  • Tech Writer

The team follows a modular, milestone-based process and is open to ecosystem collaboration.

Company Name (if applicable)

Konma Labz

Project details

High-Level Overview

To address the systemic challenges in deploying scalable, transparent and ethically verifiable AGI, we propose the development of a proof-of-concept hardware-software stack that enables symbolic AGI agents to run on open, low-power and trustless platforms — without reliance on centralized infrastructure or opaque cloud systems.

Our solution brings together three transformative elements:

Open AGI Hardware Based on RISC-V

We start by leveraging RISC-V, an open instruction set architecture (ISA), to design lightweight AGI execution environments that can operate on edge devices — from embedded microcontrollers to SoCs and FPGA-based development boards. This breaks the dependency on proprietary GPUs or ASICs and enables AGI logic to be distributed and embedded in a wide range of real-world environments (e.g., mobile robots, smart sensors, autonomous agents).

Symbolic Reasoning Runtime

On top of the RISC-V stack, we build a symbolic runtime engine capable of executing programs written in reasoning-oriented languages such as MeTTa, compatible with OpenCog Hyperon. This allows AGI agents to operate not just on data-driven models but to reason over goals, values and intentions — a critical capability for explainable, aligned intelligence. It supports recursive, graph-based cognition and goal-directed control loops.

Trustless Validation Using Zero-Knowledge Proofs (ZKPs)

To ensure verifiability of AGI decisions, we integrate RISC Zero, a zero-knowledge virtual machine that can produce cryptographic proofs of correct execution. This allows each AGI decision or reasoning trace to be accompanied by a proof that the agent followed its internal logic, without exposing sensitive inputs or proprietary reasoning graphs. These proofs can be checked off-chain or on-chain, enabling real-time accountability and decentralized trust.

Simulated Deployment Environment & Developer SDK

The system includes a testbed built using QEMU and Docker, which allows developers and researchers to emulate edge deployments and experiment with different agent configurations.

This solution lays the foundation for a new class of AGI agents that are:

  • Decentralized: Can operate on local or edge environments without centralized control.

  • Explainable: Use symbolic reasoning engines that support traceable decision paths.

  • Verifiable: Provide cryptographic guarantees that decisions are logically valid.

  • Accessible: Built with open standards, lowering the barrier to entry for developers and communities.

  • Ethical-by-Design: Ensures that alignment and safety constraints are not only encoded but enforceable and auditable.

This project directly supports SingularityNET’s vision of decentralized, democratized AGI by making it feasible to build trustworthy agents that can run anywhere — and be verified by anyone.

Open Source Licensing

Apache License

All components developed in this project will be released under the Apache License 2.0, a permissive open-source license that allows commercial use, modification, distribution and integration into both open and proprietary systems. This license ensures that the SingularityNET ecosystem and broader community can freely adopt, extend and build upon the outputs of this project without restrictive conditions.

Links and references

https://www.konma.io/r-d

www.skybrain.in

 

Proposal Video

Not Avaliable Yet

Check back later during the Feedback & Selection period for the RFP that is proposal is applied to.

  • Total Milestones

    4

  • Total Budget

    $80,000 USD

  • Last Updated

    27 May 2025

Milestone 1 - Architecture & Runtime Prototype

Description

Establish the foundational architecture for the AGI hardware stack. Select RISC-V hardware targets (real or emulated) and build a functional prototype of a symbolic reasoning runtime compatible with MeTTa-style logic.

Deliverables

A documented architecture design working symbolic logic interpreter (basic MeTTa or equivalent) running on QEMU/emulator or low-power RISC-V board and initial logic test cases.

Budget

$20,000 USD

Success Criterion

- Finalized architecture document (hardware + runtime + proof layer plan) - Symbolic interpreter runs on RISC-V (real/emulated) with ≥ 3 test cases - Code published in a versioned repository with initial documentation

Milestone 2 - ZK Proof Engine Integration

Description

Integrate the RISC Zero zkVM into the symbolic runtime layer enabling cryptographic proof-of-execution for AGI decisions. The focus is on verifying that an agent followed its encoded logic.

Deliverables

A prototype integration between the runtime and RISC Zero zkVM demonstrating generation of zero-knowledge proofs for logic evaluation paths e.g. A DAG traversal or rule execution.

Budget

$20,000 USD

Success Criterion

- Functional proof-of-execution output from symbolic runtime - Proof size ≤ 150KB; generation time ≤ 5 seconds (for test cases) - Public test suite demonstrating ZKP behavior with example logic

Milestone 3 - Edge SDK + Simulation Testbed

Description

Develop a QEMU/Docker-based simulation environment and an SDK that allows developers to package and deploy symbolic AGI logic to RISC-V edge targets. This includes CLI tools and integration hooks.

Deliverables

A developer SDK and CLI tool to deploy logic modules a simulation testbed with runtime and proof layer integration and example workflows for deployment and testing.

Budget

$20,000 USD

Success Criterion

- CLI tool enables logic packaging and deployment in simulation - QEMU/Docker testbed runs integrated stack (runtime + zkVM) - Documentation + 1 tutorial covering agent deployment and testing

Milestone 4 - Final Benchmarks Docs & Public Release

Description

Complete project with performance benchmarks power profiling (if hardware used) and public release of all components. Include integration demo for SingularityNET-compatible agents and onboarding materials.

Deliverables

Performance report public GitHub repository with MIT/Apache 2.0 license final codebase user docs and demo AGI agent. Optional delivery of a recorded demo or live walkthrough.

Budget

$20,000 USD

Success Criterion

- Benchmark report: symbolic ops/sec, ZKP latency, resource usage - Code, SDK, and docs publicly accessible and reproducible - Demo agent integrated with SingularityNET container/interface

Join the Discussion (0)

Expert Ratings

Reviews & Ratings

    No Reviews Avaliable

    Check back later by refreshing the page.

Welcome to our website!

Nice to meet you! If you have any question about our services, feel free to contact us.