Trusted AI inference hardware.
A full stack secured in silicon.
InferEdge™ EA-Series secure edge AI accelerators lead the portfolio — INT8/INT4 inference, compute-in-memory, and local decision-making for power-constrained, connectivity-limited deployment environments. Every product shares TrustCore™ hardware root of trust from validation to field deployment.
From endpoint to control plane, proof travels with the inference
Attested Infrastructure Inference (AII) is a proof-carrying attestation chain — from field endpoint through TrustCore™ Trust Continuum silicon to the Nelix Compute Platform. SecureGrid exports measurement attestation; InferEdge runs JBIA under ICSM; PCI bundles verify at fleet admit — resilient through intermittent power and constrained connectivity.
Four products. One platform.
SecureGrid™
Infrastructure silicon for metering and grid control. SecureGrid S-Series anchors TrustCore on die — proof-carrying measurement attestation through the Trust Continuum at the boundary where field deployment reality meets revenue assurance.
Metering Use CasesInferEdge™
Attested Infrastructure Inference silicon for power-constrained edge deployment. InferEdge™ runs Joule-Bounded Inference Architecture (JBIA) — ICSM-managed completion under Energy Contract — exporting Proof-Carrying Inference at INT8/INT4 within a sub-15 W envelope.
Architecture DetailsTrustCore™
The trust anchor inside every Nelix die. TrustCore is the Trust Continuum — PUF-derived identity, continuous attestation bundles, and PCI signing at the export boundary. No attestation → no trusted admit.
Security ArchitectureNelix Compute Platform
Proof-Carrying Fleet Operations for deployed Nelix silicon. The Compute Platform verifies attestation bundles before trusted admit, ships signed OTA through TrustCore gateways, and queues PCI exports offline until the uplink returns.
Discuss DeploymentEvaluate Nelix silicon for your infrastructure
Talk to our engineering team about datasheets, FPGA evaluation, and deployment timelines.