iGEM Paris 2026 · Early-Access Cohort

Run Nextflow & Protein AI
inside your own cloud.
Burn credits, not runway.

Seed to Series B biotechs sit on $100k–$350k in AWS Activate or Google Cloud credits — yet lose months to idle GPU bills, Spot preemption crashes, 3TB AlphaFold database downloads, and $50k+ enterprise orchestrators.

Join BioFlow BYOC Waitlist → Why existing options fail View BYOC Architecture
// 01 · the problem

Built for the gap between raw cloud and $50k/yr SaaS.

Why Seed and Series A TechBio teams struggle with existing bioinformatics and AI platforms.

The SaaS Markup Trap

Hosted Bio-SaaS Locks Out Your Cloud Credits

Browser-based bioinformatics platforms run compute in their cloud account and charge 2–3x compute markups — making your $100k–$350k AWS Activate or GCP Startup credits unusable while moving sensitive genomic IP outside your VPC.

The BioFlow BYOC Solution

Your VPC. Your Credits. Turnkey Orchestration.

BioFlow deploys inside your own AWS or GCP account in under an hour. Wet-lab scientists get a clean web launcher for Nextflow and Protein AI, while your cloud credits cover 100% of the underlying Spot GPU and storage spend.

The DIY DevOps Bottleneck

PhDs Shouldn’t Babysit Spot GPUs & IAM

Wiring AWS Batch, FSx Lustre caches for 3.2TB AlphaFold databases, and Nextflow Spot preemption recovery takes weeks of senior Cloud Architect time — leaving computational biologists stuck on oversized on-demand VMs.

// 02 · architecture

1-Click Terraform deployment inside your VPC.

Zero data leaves your cloud estate. Scales to $0.00/hr idle burn the second your pipeline finishes.

  • 100% Bring-Your-Own-Cloud (AWS / GCP): Deploys directly into your VPC via Terraform. Burns your existing AWS Activate or Google Cloud Startup credits with zero compute markup.
  • Pre-Indexed 3.2TB Protein AI Cache + Spot GPUs: Run AlphaFold3, Boltz-1, RFdiffusion, ProteinMPNN & ESM-3 on Spot A10G/L40S GPUs with instant read-only weight mounting and automatic preemption checkpointing.
  • Automated 14-Day FASTQ Glacier Tiering: Cuts raw Nanopore & Illumina S3/GCS storage bills by 60%+ automatically and syncs final PDBs & QC metrics straight to Benchling and BigQuery.
bioflow-byoc // eu-west-3 (paris) & us-east-1 idle burn: $0.00/hr
$ kobix bioflow init --cloud aws --region eu-west-3
→ Provisioning Spot GPU queue (g5.xlarge / g6e.xlarge)... done
→ Mounting shared 3.2TB AlphaFold/UniRef cache... done
→ Configuring 14d raw FASTQ → Glacier Instant Retrieval... done
→ Nextflow Wave/Fusion + Benchling webhook active. ready