A US-incorporated partner: B300 and H200 clusters delivered to your US site or hosted with our Tier 1 data-center partner in the US, billed in USD, with financing for defined projects — plus the AI software that runs on them.
Haink is incorporated in Delaware and bills in USD. For a US buyer that means a domestic contracting entity, OEM-warranted hardware and no export licence for a US installation — the simplest and fastest destination we supply.
| Layer | What it covers |
|---|---|
| Compute | HGX B300 servers or GB300 NVL72 racks for new clusters; H200 where a lower entry cost fits the workload |
| Network & storage | 800G / 400G fabric, all-flash or parallel storage sized to the GPU count |
| Site | Your US data center, or US capacity with our Tier 1 partner, sized per project — GPU cluster hosting → |
| Delivery & compliance | End-user documentation, export and import paperwork, customs, delivery DDP |
| Bring-up | Installation, burn-in and acceptance before handover |
| AI software (optional) | LLM and RAG applications, document intelligence, MLOps on the same contract |
| Payment | Purchase, or financing for defined projects |
How a project runs: our procurement process → · What to prepare: what we ask before a supply request →
Yes — Haink is incorporated in Delaware, USA. US clients contract with the US entity and are billed in USD.
Yes — HGX B300 servers and GB300 NVL72 racks for new clusters, H200 where a lower entry cost fits, on Supermicro, Dell, HPE and Lenovo platforms, delivered DDP with OEM warranties.
Yes, as an option: capacity with our Tier 1 data-center partner in the US, sized per project. You can also deliver to your own facility.
Yes — document intelligence, LLM/RAG applications and ML systems are built and delivered remotely, on the same contract as the hardware if you wish.
Yes, compliance permitting. The US is the simplest destination; other countries follow their own export and import rules, and every order is screened.
We supply hardware you own rather than rent, can host it for you as an option, and build the AI software that runs on it — model, pipeline and GPUs under one contract.
Running AI on this infrastructure? Haink also builds the LLM & ML software that runs on it — model, pipeline and GPUs under one US contract.
Tell us the workload, the site and the target date — we come back with the platform, a costed bill of materials, hosting options and financing.
Not a quote? Ask a question