Menu
GigaIO | SourceCode: Gryf™

The Datacenter Has Left the Building.

GigaIO’s Gryf brings uncompromised AI inference power anywhere you can wheel it — fully independent, field-serviceable, and built for where cloud dependency is not an option.

Trusted by the World’s Most Remarkable Organizations

Argonne National Laboratory Logo
Dell Technologies Logo
Department-of-Defence_Logo-File_Greyscale_Transparent
ERICSSON_Logo-File_Greyscale_Transparent
Fujitsu_Logo-File_Greyscale_Transparent
IBM_Logo-File_Greyscale_Transparent
intel_Logo-File_Greyscale_Transparent
ITRI_Logo_File_Greyscale
Lawrence-Livermore_Logo-File_Greyscale_Transparent
MITRE_Logo-File_Greyscale_Transparent
MorganStanley_Logo-File_Greyscale_Transparent
Oxford-Nanopore-Technologies_Logo-File_Greyscale_Transparent
NTT_Logo-File_Greyscale
NVIDIA_Logo-Files_Greyscale
oakridge-national-lab-greyscale_transparent
SDSC_Logo-File_Greyscale_Transparent
Supermicro_Logo-File_Greyscale_Transparent
TACC_Logo-File_Greyscale_Transparent
University of Chicago
U-of-Wisconsin_Logo-File_Greyscale_Transparent
VISA_Logo-File_Greyscale_Transparent

AI Inferencing Where You Need It

The data has always been at the edge. The compute is finally following. 

Co-designed by GigaIO and SourceCode, Gryf™ is the world’s first carry-on suitcase-sized AI supercomputer that brings datacenter-class computing power directly to the edge.

Gryf: AI Inferencing Where You Need It

Portability & Mobility

The mission can move, and the compute moves with it:

  • Take your data and processing with you wherever you go
  • Bring datacenter-class compute to places with poor and/or expensive connectivity
  • Travel in challenging environments with a ruggedized case

Configurability & Composability

No new hardware purchase, no re-engineering — just rearrange the sleds:

  • Six different sleds for GPU, compute, storage, and networking
  • Share storage or GPUs across multiple server sleds or VMs
  • Open platform built for any accelerator (GPUs, ASICs, inference cards)

Serviceability & Availability

Dramatically improve operational availability in environments without on-site support:

  • If something fails, simply swap sleds in minutes
  • Completely self-service — no technical expertise necessary
  • No downtime even if technicians are days or weeks away
Datacenter-class Computing at the Edge

GigaIO delivers meaningful AI inferencing wherever data is created and where answers are actually needed. Gryf is the only portable edge AI platform capable of supporting the full range of data processing needs — from the latest agentic AI inferencing, to machine and deep learning applications, to full-featured ISR analytics — all in a form factor that you can take anywhere you go.

Gryf vs. Other Edge AI Platforms
PacStar MDCCurtiss-Wright VoyagerHPCAnduril Klas VoyagerHPC + GPU 4.0 EdgelineHPE EL2000 with EL220/240 Blades GryfGigaIO | SourceCode★ Winning Solution
Slots per Chassis 4 1 2 6
Max CPU Coressingle unit Intel 16-core Processor AMD EPYC™ 64-core Processor Intel Xeon ProcessorScalable to 128 cores per slot AMD EPYC™ 64-core ProcessorUp to 64 cores per slot
GPU Tier / Model Workstation-classNVIDIA RTX 5000 Quadro PacStar MDC 2.0; 1 GPU per module Workstation-classNVIDIA Ampere A4500
Voyager GPU 4.0		 Virtual Desktop FocusedNVIDIA RTX6000 Server ProOne GPU only NVIDIA RTX 6000 Server Pro 96GB, L40S 48GB or H100 NVL 94GBUp to 5 GPUs per unit
Total System Power ~250–500 W ~750 W 1,300 WDual hot-plug PSUs 2,500 WDual hot-swap PSUs
Max Storagesingle unit Up to 32 TB Up to 32 TB Up to 60 TB/blade Up to 5PBUp to 1PB per sled x 5 sleds
Best Fit For Command posts, multi-enclave C2 Expeditionary forces, tactical networks, vehicle integration Industrial edge, telecom, defense Large-scale AI inference, ISR analytics, petabyte-scale field processing
Scroll to compare →

Effortless AI Infrastructure

Deploy AI in minutes, not months. Groundbreaking platforms like ours were once thought impossible.

AI at Warp Speed

Bring processing power to where it’s needed — run AI inferencing and ML workloads where the data is being generated.

Up to 10x Compute Capability

2,500 W power capacity provides up to 10x that of competitors offering only 250-1,500 W.

Run 100 Billion Parameters

Access up to 5 datacenter-class GPUs in a single server to train larger AI models.

Up to 5PB of Storage

Add storage sleds for an almost limitless storage capacity when needed.

Use Cases

Defense & Intelligence

Powered by Gryf

Learn More

Energy and
Oil & Gas

Powered by Gryf

Learn More

Media & Broadcasting

Powered by Gryf

Learn More

Sports Analytics

Powered by Gryf

Learn More
What They Say About Us
left-logo_360x200px_idc

“GigaIO has established itself as an early leader in the AI Infrastructure market segment with a solution that is based around well-accepted industry standards and provides support for heterogeneous compute, memory, storage, and networking device types.”

Eric Burgener,
Research Vice President,
Infrastructure Systems Group, IDC

left-logo_360x200px_wwt

“Having seen the advantages that composable infrastructure provides for managing, protecting, and operating massive systems at scale, companies and countries should be looking to GigaIO’s AI fabric to help future-proof their dynamic, performance-intensive workflows.”

Earl J. DODD,
Global Practice Leader for HPC,
WWT

Ready to Get Started?
Get In Touch
Envelope

Contact Us

  • This field is for validation purposes and should be left unchanged.