Introducing Fovus: A Platform for Efficient Computational Workloads Without Cloud or Infrastructure Management
Hello everyone, I’m part of the founding team at Fovus. I’ve been working on a platform to run computational workloads with better time and cost efficiency, without having to manage infrastructure/scripts or cloud. The goal is to simplify things like capacity planning, GPU constraints, and cloud spend. Sharing a few benchmarks in case helpful:
Nextflow RNA-seq: ~0.7 hrs/sample | Sarek WGS: ~$6/sample
AlphaFold3: ~45% faster
Boltz: as low as $0.10 per prediction
OpenMM: up to 3359 ns/day
GROMACS: scaling from 20K → 1M atoms
Large-scale docking: ~2 hrs for 1,000 antibody-antigen pairs
happy to connect and compare notes.