GPU-less, Trust-less, Limit-less: Reimagining the Confidential AI Cloud
summary
Claims concrete guarantees: hardware attestation from TEEs (Intel TDX, AMD SEV, Nvidia GPU TES) with on‑chain order reports, SPCTL tooling, and open-source certification; four demos show deploy-to-API marketplace, n8n healthcare workflows, distributed vLLM across four GPU nodes, and provable multi‑party training that produces attestations on BNB chain. Results include a 319% larger training set and 75% accuracy in marketing, a drop from weeks to 1–2 hours for FDA audits, and no provider lock‑in via GPUless, verifiable execution.
AI-written summary · about
chapters (12)
- 0:00Confidential AI introduction
- 1:05Trusted execution environments
- 3:51Problems solved by confidential AI
- 8:08Super Protocol explained
- 12:45Realize case study in marketing
- 15:35BEAL case study in healthcare
- 17:43Super AI marketplace demo
- 21:40Secure AI workflows with N8N
- 26:21Distributed inference with VLLM
- 31:50Trustless cryptographic proof demo
- 33:14Multi-party model training demo
- 42:22Recap and call to action
related talks
36mSecurity for AI and Multi-Party Collaboration with Confidential Computing (Oh, and...
42mSponsored Session: Confidential Computing - New Capabilities for New Workloads
20mTrusted Hardware for AI Workloads: Extending Confidential...- 42mSponsored Session: Confidential Computing - New Capabilities for New Workloads
20mBuilding Trusted AI Factories with NVIDIA Accelerated Confidential Computing
24mStrictly Confidential: Use Cases & Demos of AI Workflows in TEEs
19mHow Can We Have Private Access to Intelligence?
18mNVIDIA Confidential Computing Attestation for Next-Generation AI