Terms of Service
Effective Date: August 2026 · GPUNexus.ai LLC
Client Ownership & Transparency
GPUNexus designs, benchmarks, and deploys private AI systems that you own entirely. We do not impose proprietary lock-in, recurring software tolls, or retain claims over your models, weights, or code.
1. Scope of Assessments & Trials
GPUNexus provides 30-minute private Model Bench trials and technical workload assessments to evaluate prompt throughput, context latency, user concurrency, and hardware sizing.
Benchmarking readouts, cost crossover calculations, and model comparisons generated on the website or during trials represent technical estimates based on empirical laboratory testing. They are provided for planning purposes and do not constitute a warranty of specific future operational throughput under differing production environments.
2. Pricing, Quotes & Statements of Work
Pricing listed on GPUNexus.ai (such as hardware reference starting figures) reflects estimated hardware component configurations and standard deployment services.
All commercial hardware builds, agent implementations, and fine-tuning programs are formalized through a dedicated written Statement of Work (SOW) or purchase agreement signed by both parties.
3. Intellectual Property (IP) Ownership
You retain 100% full and exclusive ownership of all your intellectual property.
- Proprietary Code & Prompts: All application code, system prompts, ingestion pipelines, and agent tools developed for your organization belong entirely to you.
- Model Weights: All fine-tuned model weights, adapters, LoRAs, and embedding matrices created during your engagement remain your exclusive property.
- Data & Artifacts: All documents, training datasets, and telemetry generated on your hardware remain within your exclusive ownership and perimeter.
4. Hardware Warranties & Limitation of Liability
Hardware systems deployed by GPUNexus include factory pre-installation of open inference engines (vLLM, SGLang, Open WebUI) and comprehensive burn-in load testing prior to delivery. Physical components (GPUs, processors, motherboards, memory) are warranted directly by their original equipment manufacturers (OEMs), such as NVIDIA Corporation.
To the maximum extent permitted by applicable law, GPUNexus.ai LLC’s aggregate liability for any claims arising out of services or hardware advisory shall not exceed the total fees paid by the client under the applicable Statement of Work.
5. Governing Law
These Terms of Service and any commercial engagements shall be governed by and construed in accordance with the laws of the Commonwealth of Massachusetts, United States, without regard to its conflict of law principles.
6. Contact Information
For questions regarding these Terms of Service or to request a custom engagement agreement: