What We Build
The Ecosystem
The Mission
Through the foundation of an open source training stack, Manifold Labs enhances AI accessibility and expands individual ownership of compute resources.
The Stack
Deploy Confidential Compute
A truly open stack begins with a base of
permissionless compute infrastructure
Distributed Intel Confidential VMs
Built on a permissionless compute network of trusted execution environments.
Secure Intel TDX Encryption
Workloads use Intel TDX to secure your data in transit and at rest.
Nvidia PPCIE Confidentiality
GPUs use Nvidia PCIe Confidentiality to protect your data.
Take Back Compute Ownership
Ultra-high performance at home computing,
earn passively with idle downtime.
Distributed Intel Confidential VMs
Enterprise grade security, enabled by Nvidia confidential computing.
Earn Passively
Locally run and passively monetize your individual GPU Clusters.
Hardware Variants
Available in H100, H200, and RTX6000 options.
Deploy Open Source Models
Scale AI models with high-speed inference
on a confidential distributed stack
Open-Source Inference
Sybil is an open-source inference as a service with fast models, and a simple API.
Transparent Pricing
Scale with cost efficient inference, where you only pay for your exact usage.
Run On Confidential VMs
Deploy Open Source AI Models on Encrypted Virtual Machines.
Connect With Developers
Building Manifold
Join the Manifold Community for
Product Announcements and Support