Cisco Builds AI Defense with Codex

The detail worth flagging is the directionality here: Cisco is not just using AI to write internal scripts, it is using Codex to build and iterate the security product it sells to customers. That means OpenAI's inference costs are now embedded inside Cisco's own product margins, creating a cost-pass-through dynamic that neither company has publicly quantified.

That cost structure matters more than usual given what we covered the same day: OpenAI's Q1 2026 financials showed the company burning $1.22 for every dollar earned even after stripping out stock-based compensation. Enterprise deals like the Cisco arrangement are presumably part of OpenAI's path toward closing that gap, but the math only works if Cisco-style deployments generate enough volume at margins that offset inference overhead. Right now there is no public signal on whether this contract is structured as consumption-based pricing or a flat enterprise license, and that distinction determines whether it helps or worsens OpenAI's unit economics.

Watch whether Cisco discloses AI Defense's gross margin profile in its next earnings call. If inference costs are being absorbed rather than passed through, that would confirm the deal is a reference-customer arrangement priced for visibility rather than profitability.