Local AI or cloud: choose by task.

Execution location is an architecture decision, not a privacy label. Start with the task, data and service level, then compare the full cost of the available options.

By ARCKONE
  • Evaluate each task and its data flows.
  • Include hardware, energy, operations and replacement in local TCO.
  • Keep a hybrid option when tasks have different constraints.

Decide task by task

Field extraction, document search, translation, vision, image generation and general assistants have different needs. Build a representative test set, define quality and latency thresholds, then compare actual models.

Model size alone is not enough. Context, GPU memory, concurrency, quantisation and serving software affect performance. Hardware should follow this evaluation.

Decide task by task
CriterionQuestions
DataWhat enters, where does it flow and what must be retained?
QualityWhat error rate is acceptable and who reviews?
LoadHow many users, requests and peaks?
ContinuityWhat latency and downtime are tolerable?
ChangeHow often must the model be replaced?

Local does not automatically mean private or compliant

An on-premise server can reduce some transfers, but data can still leave through logs, telemetry, backups, updates, connected tools or support access. Cloud services may offer useful contractual and technical controls too. Map the real flow.

Ollama offers a local-only mode and listens only on the local interface by default, but you must manage those settings. GDPR still requires suitable technical and organisational measures. Owning hardware guarantees neither privacy nor compliance.

  • Network, authentication and role-based access
  • Logs, telemetry and retention periods
  • Backups, encryption and tested recovery
  • Maintenance access and external dependencies

Local cost goes beyond the GPU price

Local TCO includes the server, the rest of the system, energy, cooling, warranty, installation, operations, patching, monitoring, backups, outages and replacement. Cloud TCO includes usage, storage, network, managed services and pricing risk.

The NVIDIA RTX PRO 6000 Blackwell Workstation Edition, with 96 GB of GDDR7 memory, is a high-end option for some models and professional workloads. It is not a minimum or a default. Tests and budget drive sizing.

Hybrid is often a reasonable architecture

Sensitive documents and stable tasks can stay on controlled infrastructure, while the cloud handles peaks or missing capabilities. Routing must be explicit: data category, allowed model, logging, fallback and human review.

Servers such as vLLM expose APIs compatible with common clients, making backend changes easier. Compatible does not mean interchangeable: test parameters, outputs, tools and limits for every model.

Check the model licence too

Open-weight does not automatically mean open source or unrestricted commercial use. Check the model card, licence, use restrictions and redistribution duties for the exact version you deploy.

Frequently asked questions

Does local AI automatically protect data?

No. It can reduce transfers when network, logs, telemetry, backups and support are configured correctly. The complete architecture still needs review.

Is local cheaper?

It depends on load, hardware, energy, maintenance and replacement. Compare full cost over the same period and at the same service level.

Can we keep the same API?

Local servers can offer APIs compatible with common standards. That layer reduces integration work, but each model and server retains different capabilities and settings.

Turn the idea into a process.

ARCKONE can analyse the process, test the assumption and build a measurable first version.

Assess my architecture