{"resourceId":"prisma-airs-october-2026-fedramp-moderate-ai-runtime-security-and-streaming-inspection-20261029","versions":[{"version":"external-ec21175b18fc743067a484c1436ce0fc2f36fadd383467b403d7d799894b2546","resource":{"id":"prisma-airs-october-2026-fedramp-moderate-ai-runtime-security-and-streaming-inspection-20261029","title":"Prisma AIRS October 2026 release notes: AI Runtime Security API Intercept available in FedRAMP Moderate with a 15% credit surcharge, inline inspection of streamed LLM responses, and AI red-teaming client without Kubernetes","organization":"Palo Alto Networks (Prisma AIRS)","sector":"AI security / runtime protection for LLM applications","geography":"United States (FedRAMP) and global","publishedAt":"Palo Alto Networks TechDocs, October 2026 new features","publicationDate":null,"eventDate":null,"sourceName":"Palo Alto Networks TechDocs","sourceLabel":"New Features - Prisma AIRS - October 2026","sourceUrl":"https://docs.paloaltonetworks.com/prisma-airs/new-features/by-date/prisma-airs/october-2026","evidenceClass":"vendor-claim","outcomeClass":"emerging","topics":["data-security","infrastructure","governance-procurement"],"finding":"Palo Alto Networks documents that AI Runtime Security (API Intercept) now runs in FedRAMP Moderate environments with stated functional parity: prompt injection defense, DLP, toxic content filtering, malicious code and URL blocking, contextual grounding and SQL guardrails. Selecting FedRAMP Moderate in a deployment profile provisions it in certified infrastructure and applies a standard 15% Flex Credit surcharge, with connected services such as Strata Cloud Manager, Panorama, logging, DLP and WildFire inside the boundary. Other October features include inline inspection of streamed LLM responses through the firewall, a standalone binary and Docker Compose option for the AI Red Teaming network channel client, and CLARA-driven migration of AWS and Azure native firewall policies.","sledRelevance":"State agencies and universities that require FedRAMP or StateRAMP-aligned services for AI applications now have a documented government-boundary option for LLM guardrails, with a clear price premium to budget.","evidence":"Opened and read the October 2026 new-features page in full.","architectureImplications":"API Intercept requires applications to call the service; streaming support applies to Network Intercept inline deployments. Red-teaming can now reach private AI targets without Kubernetes.","governanceImplications":"FedRAMP Moderate status is a vendor statement in release notes; agencies should confirm authorization scope on the FedRAMP marketplace and whether StateRAMP or TX-RAMP equivalency applies.","securityPrivacyImplications":"DLP and prompt-injection controls are claimed at parity; no detection-rate data is given. Independent research (see the prompt-injection detector study in this edition) shows detector performance varies widely by deployment.","caveats":"Vendor documentation with no performance data, customer examples or authorization package details. FedRAMP is federal; state recognition varies.","streamIds":["palo-alto-networks"],"roles":{"sales":"Sales framing: Palo Alto Networks now documents its AI runtime guardrail service inside a FedRAMP Moderate boundary, at a 15% credit premium. The customer problem for state agencies and universities is meeting cloud-authorization requirements for AI applications. A bounded engagement could compare authorization scope and cost against agency requirements. The source cannot support claims about detection effectiveness, state authorization equivalence or total cost, and the surcharge should be disclosed early.","engineering":"Engineering framing: API Intercept means each AI application must call the service, while streaming inspection applies only to inline Network Intercept. Check whether the agency’s required logging, DLP and management services stay inside the FedRAMP boundary as stated. For proof of value, test prompt-injection and data-leak policies on the agency’s own application traffic, measure false positives and latency on streamed responses, and confirm the authorization listing independently.","delivery":"Delivery framing: delivering AI runtime guardrails means building an intake path so program teams register AI applications and route calls through the service. The security office owns policy; application teams own integration. Acceptance should include measured block and false-positive rates on agency traffic and confirmed authorization scope. Risks include cost overruns from the surcharge, guardrails tuned on demo data, and assuming federal authorization satisfies state requirements."},"roleDetails":{"sales":{"customerProblem":"Agencies need AI application guardrails inside an authorized cloud boundary and a predictable cost.","stakeholders":["State CISOs","University security officers","AI program leads","Procurement officers"],"discoveryQuestions":["Which authorization (FedRAMP, StateRAMP, TX-RAMP) does policy require?","How many AI applications would call the guardrail API?","Has the 15% surcharge been budgeted?"],"valueHypothesis":"Interpretation: an authorized guardrail service could shorten approvals for AI apps; benefit unmeasured.","potentialEngagement":"A requirements and cost comparison for AI runtime guardrails within authorized boundaries.","unsupportedClaims":["Detection effectiveness","State authorization equivalence","Total cost of ownership"]},"engineering":{"fit":"Fits agencies building LLM applications that can call an API or route through inline firewalls.","architectureIntegration":"API Intercept integration per application; Network Intercept inline for streaming; connected SCM, logging and DLP inside boundary.","prerequisites":["FedRAMP option enabled in deployment profile","Application integration work","Policy definitions"],"constraints":["Authorization scope must be independently confirmed","Surcharge applies","No published efficacy data"],"security":"Validate that logs and prompts stay in the authorized boundary and that DLP policy matches data classification.","proofOfValue":["Block and false-positive rate on agency traffic","Streaming latency measurement","Marketplace authorization check"]},"delivery":{"work":"Stand up an AI application intake and guardrail integration process.","dependencies":["Authorization confirmation","Budget approval","Application team capacity"],"ownership":"Security office owns policy; application owners own integration.","skillsAdoption":"Teams need guidance on tuning guardrail policies and reading results.","governanceCheckpoints":["Authorization review","Data classification sign-off","Pilot results review"],"acceptanceCriteria":["Proposed: measured false-positive rate below an agreed threshold","Proposed: all registered AI apps routed through guardrails"],"risks":["Cost overrun","Over-trust in untested detection","Authorization mismatch"]}},"retrievedAt":"2026-10-30T03:06:16Z","enrichedAt":"2026-10-30T03:12:42Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Simpler red-teaming client formats lower the skill barrier for teams without container platforms.","procurementImplications":"The 15% Flex Credit surcharge for FedRAMP must be included in cost models and cooperative contract quotes.","operatingModelImplications":"Security teams would own guardrail policy for AI apps built by program offices, requiring a shared intake process.","sourceVerification":{"openedUrl":"https://docs.paloaltonetworks.com/prisma-airs/new-features/by-date/prisma-airs/october-2026","referenceExcerpt":"applies a standard 15% Flex Credit surcharge to cover compliance operational overhead.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}