Your private AI cloud — an agent for every employee
In-country AI on your own GPUs. Data never leaves the network — and it's one line in the enterprise contract your customers already have with you.
Sovereign agents for the enterprise
The same agent that acts for consumers — hardened for business & government, on the telco's own GPUs. It works, it doesn't just answer.
One agent per employee · data never leaves the network · every action audited & metered.
Bundle Neo with your enterprise stack
The telco already sells these. Neo is the AI layer on top — one sovereign stack, one contract, one invoice.
One sovereign stack · one contract · one invoice · zero data egress.
Usage by model · last 30 days
| Model | Requests | Input tokens | Output tokens | Cost |
|---|---|---|---|---|
| deepseek | 8 | 48,227,976 | 28,979,768 | $23.65 |
| gpt | 8 | 9,960,048 | 27,544,784 | $300.35 |
| qwen | 9 | 41,767,388 | 26,008,415 | $20.85 |
| gemini | 10 | 17,264,396 | 51,155,293 | $276.49 |
| claude | 8 | 7,184,088 | 22,780,872 | $363.27 |
Models in your private cloud
Open-weight models running in-country on your carrier's GPUs — your data never leaves the network.
Quick start
OpenAI-compatible — point any SDK at your gateway.
curl https://api.<carrier>.ai/v1/chat/completions \
-H "Authorization: Bearer $NEO_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"deepseek",
"messages":[{"role":"user","content":"Hello"}]}'Create a key on the API Keys tab. Every token is metered and billed to your contract.
Why the carrier wins
Sell GPUs at several times commodity rental margin.
Sovereign, in-country — the sale foreign clouds can't make.
Your brand, your contract, bundled into the enterprise stack.