TEE token-transfer privacy, now with DWB and BTBE to hide storage patterns instead of just trusting the enclave and hoping nobody notices.
trending13
01 02 Can remote attestation actually prove you’re talking to the TEE, or just to a relay that forwards the traffic?03 Zero-knowledge, FHE, MPC, and TEEs are now product infra in 2026, so the real choice is which trust model you can stomach.04 A PBS repeated-game model showing reputation mostly fails to stop builder frontrunning, so honesty needs TEEs or encrypted flow.05 Can control-flow attestation catch bending attacks by replaying the path symbolically instead of enumerating it?06 Paper on Hema, a formally verified mutual attestation protocol for same or mixed TEEs, since one attestation stack was not enough.07 End-to-end confidential AI for CPU+GPU TEEs, with attestation and benchmarks, not the usual single-box secure inference demo.08 Confidential-computing attestation only proves enclave integrity; this piece says StageX-style reproducible bootstraps are the missing pr...09 How do you get FHE LLM serving out of the latency swamp? Split linear ops to CKKS, keep the messy bits in a TEE.10 Can an attested TEE router keep LLM prompts secret while the host still does auth, routing, and scheduling?11 End-to-end sovereign AI in TEEs, with OpenShift confidential containers, PETs, FHE, PQC key exchange, and auditing glued on.12 Two-Way Confidential VMs (2cVM): Collaborative Confidential Computing for Mutually Distrustful Parties arxiv.orgPaper on 2cVM, a SEV-SNP plus WebAssembly confidential VM stack for mutually distrustful parties, because one sandbox apparently wasn't e...13 A Hybrid Edge-Cloud Architecture for Low-Latency Entitlement Verification in Resource-Constrained Devices arxiv.orgHybrid edge-cloud entitlement checks with prewarmed secure caches and TEE isolation; auth, but for devices that can’t afford auth.