Common questions.
What type of organizations do you work with? +
We work exclusively with institutions — enterprises, research organizations, government agencies, and defense contractors — that require dedicated, sovereign AI infrastructure under a formal multi-year engagement. We do not serve individual developers or public cloud workloads.
What is a Take-or-Pay structure? +
A Take-or-Pay agreement means the client commits to a defined compute allocation for the full contract term and pays the agreed amount whether or not the capacity is fully utilized. This structure enables Verde Compute to procure, build, and reserve dedicated infrastructure exclusively for your organization — designed to ensure dedicated availability and isolation per engagement.
What is a Delaware SPV and why does it matter? +
Each engagement is intended to be structured through a dedicated Delaware Special Purpose Vehicle — a legally isolated entity created specifically for that client's infrastructure. This means your infrastructure, contracts, and financial flows are ring-fenced from Verde Compute's other operations and clients. It provides clean audit trails, liability isolation, and transparent governance.
Is my data and IP protected? +
Yes. All infrastructure is physically isolated — dedicated cages with biometric access control, no shared hardware. BYOK (Bring Your Own Key) encryption at rest ensures Verde Compute has zero access to your data. A mutual NDA is executed before any technical discussions begin, and all engagement terms prohibit disclosure of client identity or workload details.
What hardware does Verde Compute use? +
All deployments are built exclusively on NVIDIA Blackwell architecture. Pod scale and exact configuration are confirmed in writing at engagement — we do not offer legacy hardware at any price point. The NVIDIA AI Enterprise software stack — including full licenses — is included in every engagement. There is no mix of hardware generations within a client's allocation.
What SLA does Verde Compute provide? +
Verde Compute operates under a formally documented Service Level Agreement with defined uptime targets, response times, and credit structures. Specific SLA terms are disclosed post-NDA as part of the technical infrastructure brief. All SLA commitments are backed by contractual credit obligations, not discretionary remedies.
Where is the infrastructure located? +
Verde Compute operates within top-tier, carrier-neutral Tier-3 equivalent data centers in the Pacific Northwest, United States — powered by 100% renewable energy. Specific facility details are disclosed under NDA. All data remains within United States jurisdiction.
How long does the engagement process take? +
From submission to initial response is typically 48 hours. Typical sequence: Submission → 48h review → NDA execution via DocuSign (typically within 5 business days) → Technical brief delivery (3–5 days) → LOI negotiation (1–2 weeks) → MSA execution (2–4 weeks). Total submission-to-MSA timeline: typically 4–8 weeks. Verde Compute does not impose artificial urgency — each engagement proceeds at the client’s institutional pace.
What insurance coverage does Verde Compute carry? +
Verde Compute engagements are structured around a 20-policy institutional insurance consortium covering the full engagement term. Key layers include Cyber Liability, Commercial General Liability, Property & Equipment, Business Interruption (full-term indemnity period aligned to full contract duration), Errors & Omissions, and Directors & Officers coverage. Clients are named Additional Insured on key policy layers — meaning clients hold direct rights under the insurance contracts, not merely contractual claims against Verde. A Certificate of Insurance (COI) is provided at MSA execution.
Are you HIPAA compliant? Can you serve healthcare organizations? +
Yes. Verde Compute operates as a HIPAA Business Associate for healthcare engagements. A HIPAA BAA is executed prior to go-live for any client whose AI workloads may involve Protected Health Information. Verde never accesses, views, or processes PHI directly — client data remains entirely within the client's dedicated hardware enclave. Healthcare clients also receive NVIDIA BioNeMo (protein structure prediction, molecular generation, drug-target binding) and the FLARE Federated Learning framework (multi-site training without sharing raw patient data) as standard stack components. SOC 2 Type II controls overlap substantially with HIPAA technical safeguard requirements.
What happens at the end of the engagement term? +
At Month 30, Verde initiates a formal Month 37 planning conversation. Clients choose between two paths: (1) Evergreen — Verde upgrades to next-generation hardware (e.g., NVIDIA Rubin architecture) at a new fixed price disclosed at Month 30. Same dedicated team, same API endpoints, no migration, zero downtime. (2) Hardware Ownership — purchase the cluster at independently appraised fair market value, with 90 days free transition support and an optional Continuity MSA. Verde is the only dedicated AI infrastructure provider offering a genuine ownership path at term-end — cloud providers give you nothing at Month 37.