New TEE.Fail Side-Channel Attack Extracts Secrets from Intel and AMD DDR5 Secure Enclaves

A group of academic researchers from Georgia Tech, Purdue University and Synkhronix have developed a side-channel attack named TEE[.]Fail that enables the extraction of secrets from the trusted execution environments in Intel and AMD processors, including Intel SGX and TDX and AMD SEV-SNP with Ciphertext Hiding. The attack uses an interposition device built from off-the-shelf equipment, costing under $1,000, to physically inspect memory traffic inside a DDR5 server.

The researchers claim this method allows cryptographic keys from Intel TDX and AMD SEV-SNP with Ciphertext Hiding to be extracted, and in some cases secret attestation keys from fully updated machines in trusted status. Beyond CPU TEEs, they show attestation keys could be used to compromise Nvidia’s GPU Confidential Computing, potentially letting attackers run AI workloads without TEEs protections.

The study notes that AES-XTS encryption is deterministic and therefore not sufficient to prevent such bus interposition attacks, and OpenSSL’s cryptographic code remained vulnerable in their tests. In response, AMD said it has no plans to provide mitigations since physical vector attacks are out of scope for SEV-SNP, while Intel noted that TEE[.]Fail does not change its previous out-of-scope stance for these physical attacks.