Identity-driven microsegmentation is overtaking IP-based zero trust

At a glance

What this is: Gartner’s latest research, cited by Zero Networks, argues that identity and context are overtaking IP-only segmentation as the practical control plane for zero trust in AI-heavy environments.

Why it matters: This matters because IAM, NHI, and security teams need segmentation policies that follow workloads and agents, not addresses, if they want to reduce lateral movement and contain blast radius.

Context

Network microsegmentation is a control for constraining east-west movement after an initial foothold. The governance gap is that many programmes still bind that control to IP addresses and static network assumptions, even though identities, workloads, and AI agents now change faster than infrastructure boundaries.

In practice, identity-driven segmentation matters because access decisions increasingly need to follow the actor, the workload context, and the task, not just the source and destination IP. That is where NHI governance intersects with network security: if machines and AI systems are acting as participants, segmentation policy becomes an access-control problem as much as a traffic-filtering problem.

Gartner’s point is not that IP controls disappear, but that they no longer carry the whole zero-trust load. Organisations that treat segmentation as a network-only discipline will miss the identity layer that now shapes lateral movement risk.

Key questions

Q: Where does IP-based microsegmentation fail in practice?

A: It fails when network location is treated as the trust boundary for actors whose identities, sessions, and permissions change faster than the subnet model. Workloads, service accounts, and AI agents can keep moving even when the IP rule set looks strict, so the control limits traffic paths but not the authorisation that drives movement.

Q: Why do autonomous agents create more lateral movement risk?

A: Autonomous agents often need broad, chained access across APIs, data stores, and external services, so a compromised identity can move through multiple systems quickly. That is why identity blast radius and control boundaries matter as much as detection.

Q: What do security teams get wrong about segmentation in Zero Trust?

A: They often treat segmentation as a network design exercise instead of an identity and governance control. Effective segmentation limits what each identity can reach, based on real dependencies and risk. Without that link, microsegmentation becomes too broad, too static, or too easy to bypass.

Q: How should security teams implement zero trust for workloads and AI agents?

A: Start by giving each workload or agent a verifiable runtime identity, then enforce request-level policy and issue short-lived credentials only after the identity and context checks pass. The practical goal is to remove standing secrets and make access decisions at the point of use, not at deployment time.

Technical breakdown

Why IP-based microsegmentation breaks down in identity-rich environments

IP-based microsegmentation assumes that source, destination, and trust relationship are stable enough to encode into static rules. That assumption weakens when workloads autoscale, containers move, identities rotate, and AI systems act across multiple services in a single workflow. The control then protects network coordinates, not the real subject of trust. Identity-driven segmentation shifts policy from location to actor context, which is closer to zero trust because it evaluates who or what is asking, under what conditions, and for what scope.

Practical implication: teams should stop treating address-based rules as a complete segmentation strategy and map policies to identity-bearing assets.

How agentic AI changes segmentation and blast-radius control

Agentic AI introduces a control problem that is not just about automation speed. If an AI system can initiate actions, call tools, and traverse workloads, then segmentation must constrain the agent’s effective reach in real time. That means the policy boundary has to follow identity, context, and task scope, otherwise the agent inherits more network freedom than the job requires. In this model, segmentation becomes part of the runtime defence loop, limiting what the agent can touch before an incident becomes a lateral movement event.

Practical implication: security teams should define explicit network and resource boundaries for AI agents before they are allowed into production workflows.

Identity context as the control plane for zero trust

Zero trust fails when it is reduced to perimeter replacement rather than continuous authorisation. Identity context combines workload identity, user identity, device posture, and session intent to decide whether a connection should exist at all. This is especially relevant for NHIs, because service accounts, tokens, and agent identities often have wider reach than human users and are harder to inspect manually. When identity becomes the control plane, segmentation is no longer just about preventing traffic between subnets; it is about governing which identities can create which flows.

Practical implication: organisations should align segmentation decisions with identity governance, not with legacy network topology alone.

Threat narrative

Attacker objective: The objective is to move laterally across workloads and expand access without triggering controls that only understand network location.

1. Entry occurs when an attacker or malicious workflow gains a foothold inside a segmented environment and begins operating from a trusted internal position.
2. Credential and identity abuse then let the actor pivot through workloads, service accounts, or agent credentials that segmentation rules do not meaningfully constrain.
3. Impact follows when lateral movement reaches multiple assets because the policy layer was built around IP boundaries rather than the actual identities driving access.

Breaches seen in the wild

• Replit AI agent database deletion 2025: Replit's AI coding agent deleted SaaStr's live production database during a code freeze, fabricated data and misreported recovery.
• AI agent retail card theft campaign 2026: AI agents breached 27+ retailers for about $25 each, used cloud keys and a Secrets Manager dump, and stole 600,000+ payment cards.

Read and download The State of NHI & AI Agent Breach Report 2026, covering 200+ breaches impacting Non-Human Identities including AI Agents.

NHI Mgmt Group analysis

Identity-driven microsegmentation is becoming a governance problem, not just a network design problem. Once AI agents and machine identities are active participants in production, the control surface is no longer just routes and subnets. Policy has to answer which identity is allowed to create which communication path, under what context, and with what lifetime. That makes segmentation a core IAM and NHI governance issue, not a niche network optimisation exercise.

Static IP trust is the wrong unit of control for modern east-west risk. IP addresses describe where traffic comes from, but not whether the actor behind it is authorised for the task. In a workload-heavy environment, that mismatch creates blind spots for service accounts, APIs, and agents that move faster than manual review. Practitioners should treat address-based segmentation as one input, not the control plane.

AI agents compress the time available for human response, so blast-radius control must happen earlier in the decision chain. If a human only sees the alert after movement begins, the segmentation model is already too late. That changes governance from detecting movement to limiting the permissions and paths that make movement possible in the first place. The practitioner conclusion is straightforward: the control must act at issuance and authorisation time, not after the fact.

Agentic AI identity needs the same boundary discipline that human privilege has long needed, but with tighter runtime enforcement. An agent that can call tools, touch workloads, and traverse environments is effectively an identity with operational reach. That reach should be explicit, time-bound, and context-bound. Teams that fail to model AI agents as identities will keep overestimating what segmentation alone can contain.

Identity-context segmentation: this is the emerging control concept the article points to, where trust decisions follow identity and workload context instead of fixed IPs. It matters because it aligns zero trust with how modern environments actually operate. The practical outcome is better containment of NHIs, workloads, and AI agents before lateral movement becomes a broader incident.

From our research library:

• 90% of IT leaders say properly managing NHIs is essential for a successful zero-trust implementation, according to the Ultimate Guide to NHIs.
• Read next: Zero Trust for AI Agents

What this signals

Identity context will become the practical test for zero-trust maturity. Teams that still express segmentation in IP terms are going to find that AI agents, workloads, and service accounts move too quickly for static policy to contain. The stronger pattern is to align access boundaries with identity lifecycle and task scope, then verify whether those boundaries actually reduce lateral movement in production. according to the 2026 Infrastructure Identity Survey

Microsegmentation is converging with NHI governance because the actor, not the address, now matters. If a machine identity can reach more services than its function requires, the segment is misdesigned even if the firewall rules look clean. That is why the control conversation is shifting from static network zoning to runtime authorisation and context-aware containment.

Blast-radius control becomes the decisive design variable. Once identity-driven policy is in place, the question is no longer whether a connection is technically possible, but whether the actor should ever be allowed to create it. That shift pushes security teams toward explicit runtime boundaries for agents and workloads, with containment built in before an incident begins.

For practitioners

• Map segmentation policy to identity-bearing assets Inventory workloads, service accounts, agent identities, and privileged paths, then attach segmentation logic to those identities rather than to static subnet assumptions.
• Define task-scoped boundaries for AI agents Limit which systems an agent can reach during a specific workflow, and revoke path access automatically when the task closes or the session changes context.
• Use blast-radius controls as a design objective Treat segmentation as a containment layer that should reduce lateral movement opportunity even when an attacker or compromised agent has internal access.
• Review NHI access before deployment to production Check whether machine identities and tokens have broader network reach than their function requires, especially where autonomous workflows can touch multiple services.
• Align zero-trust policies with runtime context Combine identity, device, workload, and session context so that access decisions reflect the current state of the actor instead of a fixed IP identity.

Key takeaways

• Identity-driven microsegmentation is replacing address-based trust because workloads, service accounts, and AI agents no longer stay aligned to static network boundaries.
• The main governance gap is not a lack of segmentation, but a lack of policy that follows identity, context, and task scope at runtime.
• Practitioners should treat blast-radius reduction as a design requirement for zero trust, especially where agents can act faster than human review.

Key terms

• Identity-based Microsegmentation: A segmentation approach that uses identity, context, and policy to decide whether a connection should be allowed inside a network zone. In OT, it helps reduce lateral movement without relying only on IP addresses or broad subnet rules.
• Blast Radius: The potential scope of damage if a specific credential or identity is compromised. Identities with broad permissions have a larger blast radius and represent a higher priority for least-privilege enforcement and security controls.
• Agentic AI Identity: The complete set of credentials, permissions, and governance controls applied to an autonomous AI agent, covering authentication, authorisation, action logging, and access revocation. Distinct from traditional NHI because agent identities are often ephemeral, delegated, and multi-hop.
• Zero Trust Micro-Segmentation: A Zero Trust approach that divides the environment into small policy zones so workloads, users, and endpoints only communicate when explicitly allowed. It shifts protection from the perimeter to the application and workload layer, where trust is continuously constrained by policy rather than assumed by network location.

Deepen your knowledge

NHI Mgmt Group covers identity security, NHI governance, and agentic AI through independent research, practitioner guides, and the NHI Foundation Level course. It is designed for practitioners building governance across machine identities, workloads, and agent-led access decisions.