Here’s what’s been happening at CETI AI over the past two weeks:
Intelligent Compute Fabric (ICF) Development
We’ve continued refining the ICF architecture with two major advancements:
-
Event-Based Database System
We’ve transitioned from SQL databases to an append-only log database (inspired by blockchain) using CurrentDB. This upgrade enhances:-
Concurrency handling
-
Real-time tracking of hardware usage
-
Transparent, immutable billing foundations
-
-
Federated Kubernetes Autoscaling
Curtis demoed how the ICF can now remotely deploy workloads to clusters—even behind firewalls—by using outbound-only connections. This enables:-
Enlisting clusters globally without custom networking
-
Permissionless infrastructure growth
-
Kubernetes-native control with Helm
-
Why This Matters
Previously, hardware had to be inside CETI’s main cluster. Now, any compute resource globally—rented, bare-metal, or community-contributed—can be added with ease. We’re building toward a truly decentralized, elastic AI compute layer.What’s Coming Next
-
August 2025: Finalizing the enlistment flow to allow anyone to connect their GPUs/clusters
-
Sept–Oct 2025: Billing system rollout with provable, auditable billing and crypto payouts via Wire Network
-
Q4 2025: Demand integrations (e.g. Psyc, Sire, Alpaca, Morpheus, Dula), DevOps-as-a-Service offerings, and autoscaling monetization pathways
Ecosystem & Business Development
-
Conversations continue with partners like Dula, Morpheus, and Alpaca to support scalable AI workloads.
-
CETI is increasingly seen as the DevOps backbone for builders in decentralized AI, enabling infrastructure abstraction for inference and training demand sources.
-
Partnerships will route compute from unused global GPUs to high-traffic AI services with automated monetization.
- Held our monthly town hall with the community, with transcripts provided for our international holders to follow along with development.
