Dell and NTT Data are offering compact server systems that enable Kenyan businesses to run selected artificial‑intelligence applications within their own premises, avoiding the need for large data‑centre facilities or constant cloud connectivity.

NTT Data’s general manager for technology solutions, Andrew Ngunjiri, says firms can start with a modest deployment rather than building the extensive infrastructure traditionally required for enterprise AI.

The offering, described as “AI in a box”, runs smaller language models on equipment placed close to the data source, allowing tasks such as document summarisation, information extraction and internal knowledge search to be performed locally.

Local deployment can help organisations that handle confidential customer records, financial data or contracts keep processing within company‑controlled infrastructure, reducing exposure to external cloud providers.

Keeping AI on‑site may also improve response times where network latency is an issue, though actual performance depends on the server hardware, model and workload.

Kenya’s Data Protection Act, 2019 governs personal‑data processing and sets conditions for cross‑border transfers, but does not mandate that all business data remain in the country; locally hosted AI can assist firms in meeting their chosen data‑handling controls.

Cost considerations extend beyond the purchase price; businesses must budget for electricity, maintenance, software licences, security updates and staff to manage the servers, while specialised hardware may require additional investment.

For occasional AI use, a cloud subscription may still be cheaper, but firms with high‑volume, repetitive document processing—such as invoice handling or internal record retrieval—could achieve better long‑term value with on‑premise servers.

The push aligns with Kenya’s National Artificial Intelligence Strategy 2025‑2030, which emphasises digital infrastructure and data governance, and could broaden AI experimentation among SMEs that lack resources for large‑scale computing.

Dell and NTT Data suggest a hybrid approach, where businesses combine local AI for sensitive or latency‑critical workloads with cloud services for tasks demanding greater compute power, allowing flexible scaling as needs evolve.