32-bit Binary Towers are 5x more hardware efficient than Mersenne31 multipliers

CPU proof generation bottlenecks are typically tackled by GPU- and FPGA-based hardware acceleration.

We discuss a GKR-based approach

Announcing collaboration with Polygon Labs on Binius-based zkVM

A major update to FRI-Binius yields better batching, faster recursion, and smaller proofs

ZKP requires high-performance compute to bring cost-effective, real-time transaction validation to web3.

In collaboration with 3MI Labs, we introduce a arithmetization-oriented hash function designed for Binius.

We report the first end-to-end, FPGA-accelerated Polygon zkEVM prover.

Ulvetanna is rebranding to Irreducible

In a new research paper, we adapt BaseFold's multivariate FRI commitments to binary towers, reducing Binius proof sizes.

We present a novel system demo accelerating elliptic curve cryptography on AMD's Versal heterogeneous compute platform.

Jim Posen and Radi Cojbasic from Irreducible on the Zero Knowledge podcast.

We demonstrate a performance-optimized approach to SNARK construction based on towers of binary fields.

We report the first fully-pipelined FPGA architecture for ZKP-friendly Merkle trees using the Poseidon hash

Our new result in error-correcting codes yields a 2x zk-SNARK speedup

We present a modular, fully-pipelined FPGA architecture for large NTT computations in the 64-bit Goldilocks finite field

Announcing our $15M Seed Round co-led by Bain Capital Crypto and Paradigm with participation from Jump Crypto

New work on lookup arguments with reduced complexity