Ask FirstNet about Sovereign AI
Sovereign AI, hosted at home
FirstCoreAI is FirstNet's AI business unit. It gives South African organisations GPU compute, open-weight models and engineering talent to run serious AI workloads, while the data stays inside the country and the bill stays in rand.
Choose your service
Sovereign AI services
- GPU & InferenceInfraAI delivers FirstCoreAI's AI Factory as a service: private inference endpoints, dedicated or shared GPU capacity and a…Explore
- Custom AI EngineeringInfrastructure alone does not fix a business problem. EngineAI is FirstCoreAI's build team: they take a use case from proof of…Explore
- Creative & Content AIStudioAI is FirstCoreAI's offering for when content itself is the workload: personalised video, voiceover, copy and design…Explore
- AI Adoption & EnablementFirstCoreAI helps your people use AI well, safely and consistently, with private assistants that run on the Sovereign AI…Explore
Who runs it
FirstCoreAI, inside FirstNet's data centre
FirstCoreAI is a business unit of FirstNet, the data-centre, connectivity and ISP arm of the First Technology Group. Its AI Factory GPU cluster sits in FirstNet's South African data centre, on networks FirstNet operates, so the team that runs your connectivity also runs the platform your AI lives on.
- Physical compute owned and operated in South Africa
- Blackwell-class GPU cluster with OpenAI-compatible APIs
- Network tuning, rack scaling and cross-connects handled by one team
- Backed by a group with over three decades in South African IT
Quick answers
What is sovereign AI?
It is AI that runs on infrastructure located in your country, operated under its laws and priced in its currency. With FirstCoreAI, inference happens on GPUs in FirstNet's South African data centre, so your data stays in South Africa.
How does FirstCoreAI relate to FirstNet?
FirstCoreAI is a business unit of FirstNet, the data-centre, connectivity and ISP arm of the First Technology Group. Its AI Factory runs inside FirstNet's data centre on FirstNet-operated networks.
Does sovereign AI help with POPIA compliance?
FirstCoreAI designs its endpoints, retention and access controls to align with POPIA obligations. Because processing happens locally, there is no cross-border transfer to reconcile for workloads that stay on the AI Factory.
Can I still use frontier models like those from overseas labs?
Yes. Sovereign does not mean isolated. Most workloads run well on open-weight models, and frontier APIs can be used where a workload justifies it. FirstCoreAI tells you plainly when a request would leave the country.
Why it matters
What sovereign AI means for you
Sovereign AI is AI that runs on infrastructure in your own country, under your own laws, and priced in your own currency. For South African businesses, it addresses three pressures that offshore AI services leave open: data residency, foreign legal reach and dollar-linked costs.
POPIA and data residency
Sending customer records, claims or call recordings to an offshore API is a cross-border transfer, with consent and breach questions attached. When inference happens in South Africa, the data never leaves and there is nothing to reconcile.
Jurisdiction and reach
Data held by foreign providers can fall under foreign law, such as the US CLOUD Act. Local, sovereign hosting means only the people you have authorised can reach your information.
Rand billing
Offshore AI is typically metered per token in dollars and exposed to exchange-rate swings. Open-weight models on local hardware turn that into predictable, rand-denominated spend that finance teams can actually forecast.
Solutions
Four solution lines, one sovereign platform
Each line works on its own, but most organisations combine two or three: compute, engineering, content and adoption, all delivered from South Africa.
GPU & Inference (InfraAI)
Private inference endpoints and GPU capacity from the AI Factory, hosted locally and billed in rand.
Custom AI Engineering (EngineAI)
RAG systems, AI agents, document and call pipelines, and fine-tuned models built for your problem.
Creative & Content AI (StudioAI)
Personalised video, voice, copy and design produced at scale and kept on-brand.
AI Adoption
Private AI assistants, training and governance so the tools are actually used.
Explore the details
The AI factory, by the numbersThese figures were measured by FirstCoreAI on the production AI Factory under real serving conditions, not…
Measured performance
The AI factory, by the numbers
These figures were measured by FirstCoreAI on the production AI Factory under real serving conditions, not taken from a vendor datasheet.
~0.6s time to first token
Median for the chat and coding models, staying under a second during concurrent bursts.
271 concurrent requests
Sustained across the cluster at the peak of the full stress run.
0 errors
Every request in the stress-test suite completed at every concurrency level tested.
6 live endpoints
Production model endpoints exposed through an OpenAI-compatible API.
Where it works, and how to startSovereign AI matters most where data is sensitive and volumes are high. FirstCoreAI is already in…
Industries and next steps
Where it works, and how to start
Sovereign AI matters most where data is sensitive and volumes are high. FirstCoreAI is already in production for financial services and insurance, retail and FMCG, and is built for public-sector and regulated workloads where residency is a legal requirement. An engagement starts with a 30-minute scoping call or a tour of the AI Factory, booked directly with FirstCoreAI.
- Financial services and insurance: claims, call archives, document extraction
- Retail and FMCG: personalisation, campaign automation, analytics, model hosting
- Public sector: workloads where foreign jurisdiction is not an option
- Start with one workload; FirstCoreAI shows it running sovereign
- Book a demo or scoping call with FirstCoreAI
How is sovereign AI billed?
Commercials are rand-denominated and workload-based, so costs are predictable and not exposed to exchange-rate movements. Pricing is discussed during scoping with FirstCoreAI.
How do we get started?
Bring one workload, such as a document pipeline, call archive, chatbot or a model you cannot deploy. Book a scoping call or AI Factory tour with FirstCoreAI and they will show it running sovereign.
From the Knowledge Hub
- What kind of custom AI solutions does FirstNet's EngineAI team build?
- What happens if FirstNet's FirstCoreAI finds AI isn't the right fit for our problem?
- Can FirstNet's EngineAI team fine-tune an AI model to our industry terminology?
- Can FirstNet's FirstCoreAI process call recordings with AI without sending them offshore?
- Where are FirstNet's EngineAI custom AI solutions deployed?
- Can we see a demo of FirstNet's StudioAI creative content service?
See it running on home soil
Your place in the stack
Sovereign AI is layer 5 of 5.
The intelligence layer sits on top of everything FirstNet runs, and can work with data from every layer below it. See the full stack →
Bring us a problem.
We’ll show you the stack that solves it.
Ask FirstNet for an instant answer, or leave your details in the chat and the right specialist will contact you.