Urgent.News

What's breaking now, across thousands of outlets.

AI

Local AI cloud project puts over 400 Nvidia GPUs to work

Stratos Lab, Ecoblox and Digital Parks Africa partner to deploy an AI cloud infrastructure platform that is expected to deliver 7.2 exaflops of computing performance.

Local AI cloud project puts over 400 Nvidia GPUs to work

South African cloud provider Stratos Lab, AI infrastructure company Ecoblox, and data centre operator Digital Parks Africa have teamed up to deploy an AI cloud platform in the country, utilizing over 400 Nvidia GPUs. The infrastructure, comprising more than 50 Nvidia B300 HGX servers from Giga Computing, is expected to deliver 7.2 exaflops of computing power and will focus on AI and high-performance computing workloads.

The project will be hosted at Digital Parks Africa's data centre, with Stratos Lab providing cloud and compute services, Ecoblox managing systems integration and deployment, and Ecoblox handling systems integration and infrastructure deployment. The goal is to give African businesses and developers access to high-performance AI computing within the continent, reducing reliance on overseas resources.

Stratos Lab CEO Chris Mostert says the deployment signifies a significant leap for regional innovation, while Ecoblox CEO Doug Makishima emphasizes the need to address AI scaling challenges. Digital Parks Africa's Menno Parsons highlights the importance of a robust physical foundation for ultra-dense AI workloads. Giga Computing will supply the GPU servers, supporting the project with its R&D resources and cost optimization.

Stratos Lab will offer the infrastructure through GPU-as-a-service, managed platform-as-a-service, and bare-metal computing services, targeting workloads like AI model training, inference, and AI-agent applications.

Written by urgent.news from ITWeb's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at itweb.co.za →

More in AI

Agentic AI Needs a New Kind of Observability, Here's Why

For the last decade, observability has been built around a fairly simple mental model: a request comes in, it moves through a handful of services, and somewhere in that path, something goes wrong.

  • Agentic AI's unpredictable nature challenges traditional observability models.
  • New observability approaches focus on capturing intermediate reasoning steps and tool choices.
  • Teams are developing custom observability solutions for agentic AI monitoring.

More from Thursday 27 August →