S1 · Personal AI Compute
Meaningful AI capability you can own locally.
A system class for developers, researchers, engineers, founders, creators, students, and serious AI users.
Pre-launch status
No Pleroma compute product is currently offered for sale. Final configuration, pricing, availability, shipping, and warranty terms will be published only after supplier economics and launch readiness are confirmed.
Current timing: To be announced. Product and Offer structured data remain disabled.
Compute architecture
Linux-forward, Arm-focused, and NVIDIA-based design direction. Final platform is not yet selected.
Runtime memory
128GB+ coherent or unified memory class target for meaningful local model execution.
Model storage
4TB baseline target with a path toward 8TB. Storage is treated as capability, not an accessory.
Interfaces
Standard development frameworks and interfaces, with optional Pleroma integration rather than forced captivity.
Design objective
Capability first.
S1 is defined by useful local AI capability, not by a logo or a speculative product name. If the launch system is a manufacturer-branded product, the product page will identify it accurately.
- Own model artifacts locally
- Run meaningful workloads without mandatory cloud dependence
- Use standard tooling and interfaces
- Burst extreme workloads remotely when that is economical
Why is there no price yet?
A planning assumption is not a commercial offer. Pleroma Works will publish one authoritative live price only after landed supplier economics, inventory, fulfillment, warranty, and launch readiness are confirmed.
See the pricing mechanism