Altius brings institutional-grade, high-performance execution layer to any blockchain, removing infrastructure complexities so teams can build at scale.

Altius is designed to optimize blockchain performance through a VM-agnostic execution framework that enhances scalability, efficiency, and cross-chain interoperability. By decoupling the execution layer from monolithic network designs, the Altius Stack integrates seamlessly with Layer 1s, Layer 2s, and application-specific chains, providing immediate performance improvements without requiring specialized hardware.

Designed with a modular architecture tailored for a multi-chain future, Altius empowers any blockchain to achieve faster transactions, reduced costs, and enhanced interoperability, futureproofing them for the next generation of on-chain applications.

Altius's core team includes former Managing Partner of Amber Group and a 14-year HFT distributed systems veteran from Hudson River Trading, and is backed by Founders Fund and Pantera Capital: https://www.bloomberg.com/news/articles/2025-02-18/hudson-river-amber-crypto-quant-vets-want-to-take-on-solana

Technical architecture

Read our whitepaper here: https://www.altiuslabs.xyz/blog/altius-whitepaper

Alpha release

We have launched an EVM implementation with the altius-revm parallel execution engine and altius-reth, achieving a 4× improvement over sequential execution and proportional gains in gas-per-second throughput.

Key highlights:

To break it down further, we have run benchmarks under various scenarios:

🔹 Complex DeFi workloads (Uniswap-style transactions)

🔹 On commodity hardware (no high-end tuning)

The results demonstrates that the Altius execution engine for Ethereum-compatible systems combines speed, determinism, and scalability across any workload, from ERC20 transfers to the most complex DeFi protocols.

Benchmark Results

Workload Legacy EVM (Sequential) Altius Parallel+SSA Improvement
ERC20 Transfers 3.38s (1,627 Mgas/s) 0.905s (6,077 Mgas/s) 4× faster
Uniswap (1k txs) 55.8ms (2,455 Mgas/s) 32.5ms (4,224 Mgas/s) 2.3× faster
KZG Precompiles 192ms (617 Mgas/s) 41ms latency (2,900 Mgas/s) 4.7x faster

Same final state root across all modes → deterministic correctness preserved