TL;DR: Retail and e-commerce collaboration can keep raw data and model artefacts protected during computation when access depends on attested runtime integrity, according to Fortanix. Confidential AI shifts governance toward evidence-based key release and auditability instead of broad trust in partner infrastructure.
Editorial analysis by NHI Mgmt Group, based on content published by Fortanix: “How Fortanix Confidential AI Helps Secure Sensitive Enterprise Data for Retail & E-commerce”.
Key questions
Q: Why do trusted execution environments matter for cross-organisation AI projects?
A: They reduce the trust surface by isolating code and data during processing, which is the stage where many collaboration models become risky.
A: They should verify the runtime through remote attestation, confirm the environment matches the approved configuration, and check that key release is tied to that proof.
Q: How should organisations share sensitive data for joint analytics without exposing raw records?
A: Use a model where the computation happens inside a verified trusted execution environment and decryption keys are released only after attestation succeeds.
Practitioner guidance
- Define the collaboration boundary first Map exactly which retailer, supplier, or logistics datasets must be protected, which party controls each key, and what must never leave the trusted execution environment.
- Require attestation before key release Make remote attestation a mandatory gate for any workload that requests access to sensitive data, model weights, or shared analytics inputs.
- Separate compute trust from data trust Ensure the infrastructure that runs the workload does not also have standing authority to decrypt the data it processes.
Bottom line: Retail collaboration becomes easier to govern when access depends on attested runtime integrity instead of broad trust in partner infrastructure.
Glossary terms: Confidential AI, Trusted Execution Environment, Remote Attestation, Attestation-Gated Key Release
Explore further
View Full Forum → | NHI Foundation Course → | Our Services → | Read the full analysis →
Confidential AI is becoming a governance pattern, not just an infrastructure feature. Retail collaboration fails when organisations are asked to trust every layer between the dataset and the computation. Confidential AI changes the decision model by making trust conditional on measured runtime integrity rather than on contractual assurances or network segmentation. That is a meaningful shift for any programme that needs to share sensitive data across organisational boundaries.
A few things that frame the scale:
- Only 5.7% of organisations have full visibility into their service accounts, according to the Ultimate Guide to NHIs.
A question worth separating out:
Q: How can teams tell whether confidential AI is actually improving data governance?
A: Look for evidence that access decisions are being made at runtime, that keys are only released to attested environments, and that audit records show what executed without exposing the underlying data. If the programme still relies mainly on policy statements, delayed sharing, or broad infrastructure trust, the control is not yet doing the work it should.
👉 Read our full editorial: Confidential AI narrows retail data-sharing risk without exposing raw data