Confidential AI Inference at Scale
Latency and Throughput Overhead of TEE-Protected Inference
TEE overhead concentrates at memory boundaries, not compute.
Lorena Sundqvist
· 11 min read
Private AI Review — Page 5
Confidential AI Inference at Scale
Multi-Party Inference with Confidential Computing
Trusted hardware lets hospitals and AI firms compute together without exposing secrets.
Kian Vance10 min read
Confidential AI Inference at Scale
Model Weight Protection Inside Confidential Inference Servers
Plaintext model weights in memory can only be protected by hardware, not policy.
Lorena Sundqvist13 min read
Confidential AI Inference at Scale
Encrypting Prompts and Responses in Transit to a TEE
Hardware-sealed enclaves protect LLM prompts from cloud operators and hypervisors during inference.
Kian Vance11 min read
Attestation and Verifiable Computing
Confidential Computing Attestation vs TLS Certificate Trust Models
Attestation proves what code is running; TLS only proves you reached the right domain.
Anouk Mercer11 min read
Attestation and Verifiable Computing
RATS Architecture and Attestation Evidence Formats
How role separation and standardized formats prevent devices from vouching for themselves.
Preston Nardone9 min read
Attestation and Verifiable Computing
Verifiable Transparency Logs for Model Provenance
Cryptographic logs can prove a model's training history wasn't altered after release.
Lorena Sundqvist18 min read
Attestation and Verifiable Computing
ARM TrustZone vs Intel SGX for Attestation Workflows
SGX roots trust in Intel's PKI; TrustZone delegates it to device vendors.
Preston Nardone12 min read
Attestation and Verifiable Computing
Attestation Token Replay and Freshness Attacks
Valid tokens become dangerous weapons when presented outside their intended time window.
Lorena Sundqvist15 min read
Attestation and Verifiable Computing
Measured Boot and Platform Integrity for AI Servers
Cryptographic boot logs catch AI server tampering that signatures alone miss.
Preston Nardone13 min read
Attestation and Verifiable Computing
in-toto Attestation for AI Supply Chains
Cryptographic verification for AI models fills the gap between documentation and actual artifacts.
Kian Vance14 min read
Private AI Use Cases and Configuration
On-Device AI Inference on Apple Silicon for Privacy-Preserving Applications
Apple's unified memory lets developers run AI models on phones without sacrificing privacy.
Preston Nardone11 min read
