Moonmath (Ingonyama)

Stage: Seed

Get updates & announcements

About

Building the Performance Layer for World-Models

Contact Details

hi@moonmath.ai

https://www.moonmath.ai

https://moonmath.ai/posts

Community Affiliations

Host

ZK Accelerate

Nov 2024

ZK Accelerate

Nov 2024

Sponsor

Members (3)

Omer Shlomovits

Founder & CEO

Otsar Shalmoni

Marketing Manager

Shira Robas

Director of Operations

Focus Areas

Projects (3)

Blaze

Blaze is a Rust library for ZK acceleration on Xilinx FPGAs.

Icicle

A GPU Library for Zero-Knowledge Acceleration

ZaKi

ZK Proving with AI Hardware

Moonmath (Ingonyama) News (3)

MoonMath (Ingonyama) shipped SGLang integration and a faster MI300X decode kernel

Two engineering posts detail SGLang support for Zro and a new MLA decode kernel that beats AMD's AITER v3 by up to 1.26x on long-context inference.

Launchzro.moonmath.ai1 month ago

3 Views

Ingonyama announced a pilot program serving open-source AI models for coding, targeting users who prioritize data privacy and performance over proprietary cloud inference services.

Jun 23, 2026 — Ingonyama announced a pilot program serving open-source AI models for coding, targeting users who prioritize data privacy and performance over proprietary cloud inference services.

Announcementx.com3 months ago

2 Views

MoonMath AI open-sourced a HIP attention kernel for AMD's MI300X GPU — using one-instruction assembly wrappers and an eight-wave pipeline — which benchmarks show outperforms AMD's…

Jun 19, 2026 — MoonMath AI open-sourced a HIP attention kernel for AMD's MI300X GPU — using one-instruction assembly wrappers and an eight-wave pipeline — which benchmarks show outperforms AMD's own AITER v3 library; code is released under MIT and a pull request is under review for SGLang.

Othergenerativedaily.com3 months ago

2 Views

All network updates