Five signals every morning across three domains: AI and frontier models, energy and electricity, corporate strategy and state-owned enterprises. Every signal must cite a source you can check yourself.

OpenAI formalises Zero Data Retention and previews Private Safety Processing for frontier API users

What changed

OpenAI publicly reaffirmed that eligible API customers can operate under Zero Data Retention—meaning no prompts or completions are stored by OpenAI—and previewed Private Safety Processing, which runs safety evaluations on encrypted data without exposing plaintext content to OpenAI infrastructure. [1]

Why it matters

Enterprise AI adoption has been gated less by model performance than by data-governance compliance: regulated sectors (finance, health, defence) require contractual certainty that inputs do not become training data. Formalising ZDR as a named, documented policy for frontier models, rather than a negotiated carve-out, turns data retention into a standard procurement checkbox rather than a bespoke legal negotiation.

Second-order effect

If ZDR becomes a baseline expectation across frontier providers, the competitive differentiator shifts from 'we don't train on your data' (a negative commitment) to the quality of the safety and audit tooling that operates within the privacy perimeter—i.e., Private Safety Processing becomes the next moat. Rivals who cannot offer equivalent encrypted safety evaluation will face enterprise procurement disqualification independent of model quality.

Action

Procurement and legal teams evaluating frontier AI contracts should update vendor scorecards this week: add a ZDR-equivalent column and a 'safety-within-privacy' capability column. Assess whether current MSAs with other providers achieve the same standard by contract or merely by policy (which is revocable).