ABD Fabric
Founding Partner Program

Prove ABD Fabric on your hardware, with your production-representative workloads. If the pilot performs against the targets we agree up front, lock founder economics and help shape the interoperability layer that breaks CUDA vendor lock-in.

Apply for the Program See How It Works →
90 days No license or runtime fees during the validation pilot.
35% Founder discount locked for the life of the initial contract term.
0 Application code changes required for CUDA workloads in the pilot.
2.19x Throughput observed vs. NVIDIA H100 in tested workloads.

Performance varies by workload, optimization depth, and hardware mix. Program success criteria are confirmed with each partner before pilot kickoff.

Capacity is scarce. Capital is stranded. Proof needs to happen at scale.

AI infrastructure demand is racing ahead of deployable capacity, while hardware incompatibility leaves too much capital locked behind one accelerator ecosystem.

ABD Fabric already runs unmodified CUDA workloads on AMD. The Founding Partner Program is designed to validate functionality, performance, cost, and scalability inside real enterprise and data-center environments.

Four founder rewards for going first

$

Founder Economics

A no-cost 90-day pilot, waived runtime fees during the pilot, waived first-deployment integration, and a locked 35% discount on conversion.

⚡

Early Access

Pre-general-availability builds, private release channels, and early access to new hardware and runtime targets.

⌘

Direct Developer Access

A private channel to ABD engineering, a named deployment engineer, and priority escalation through the program.

◎

Roadmap Influence

A seat on the ABD Founding Partner Technical Council, including roadmap reviews and prioritization input.

From fit review to reference deployment

Phase 0 · Weeks 1–2

Qualify & Onboard

Mutual NDA, fit review, target workload selection, success criteria, and founder agreement.

Phase 1 · Days 1–90

Free Validation Pilot

Deploy ABD Fabric on AMD-first or existing fleet capacity. Run the workload unmodified and capture benchmarks.

Phase 2 · Days 90–180

Scale Validation

Expand node count and concurrency to prove functionality, economics, and scalability as the deployment grows.

Phase 3 · Month 6+

Convert to Founder

Lock founder pricing, participate in the Technical Council, and pursue an approved case study after successful conversion.

Founder rewards are earned through real-world validation.

The program is best suited for operators ready to test ABD Fabric against production-representative AI or HPC workloads.

Interoperability without rewriting the workload

Up to 50% CAPEX savings target on AI hardware, including capacity already owned.
Up to 30% OPEX savings target as usage scales across heterogeneous infrastructure.
40–70% Of NVIDIA cost per token observed in early modeling and testing.

Request a founding partner fit review.

Tell us about your hardware environment, workload, and validation goals. ABD will review fit and follow up to schedule a 30-minute introduction.

Prefer email? Contact greg@artificialbydesign.com.

Founding Partner Application

Fields marked with an asterisk are required.

✓ Application Received

Thank you for your interest in the ABD Fabric Founding Partner Program. The ABD team will review your submission and follow up on next steps.