The Guards

Cybersecurity

Cross-cutting force · Security across every layer

Why it matters
Security failures at any layer undermine the layers above and below. An agent with excessive permissions is a governance problem, an identity problem and a data problem at the same time.
The bottleneck
Skilled people and coherent identity. Tooling is plentiful; the ability to operate it consistently is not.
Who captures value?
Platforms with broad telemetry and a position customers are reluctant to unpick, particularly where identity and detection meet.
What could change?
Machine identity for autonomous agents is an unsolved control problem, and whoever solves it convincingly could reshape the category.

Security is not a step in the sequence. Every layer adds attack surface, from firmware in the data centre to an agent holding credentials on someone's behalf.

Each addition to the ecosystem creates something new to defend. More data in more places, more identities, more automated systems taking actions without a human reviewing each one. Security vendors are selling into organisations that are deploying AI faster than they are securing it, which is a durable commercial position and an uncomfortable fact.

SECURITY PERIMETERChips and hardwareFirmware and supply chain attacks reach below the operating system, where most security tools cannot see.Data centresPhysical access, shared tenancy and misconfigured cloud storage remain among the most common causes of large breaches.NetworksTraffic between GPUs, buildings and regions has to be encrypted and segmented without slowing training to a crawl.Models and dataTraining data can be poisoned, prompts can be injected, and model weights are valuable enough to be worth stealing outright.ApplicationsEvery AI feature bolted onto existing software is a new door into the data behind it.Autonomous agentsAgents act on a user's behalf with real permissions. A compromised agent does not leak data, it takes actions.Each new box is a new way in.

Tap or hover a box to see what it does

Every layer added to the stack is another surface to defend. Security wraps the whole thing rather than sitting on top of it.

Endpoint and cloud protection

Detecting and responding to activity on devices and in cloud environments, where misconfiguration remains one of the most common routes in.

publicCRWD

CrowdStrike

What it does

Provides cloud delivered endpoint detection and response, threat intelligence and related security modules.

Why it matters

Widely deployed across large enterprises, with a sensor footprint that improves detection as it grows. Agents acting autonomously on corporate systems make endpoint telemetry more valuable, not less.

What could go wrong

Platform consolidation pressure, and the reputational cost of any incident affecting a component installed that widely.

Position in the ecosystem
publicS

SentinelOne

What it does

Provides endpoint and cloud security with automated detection and response.

Why it matters

A credible alternative in a category where buyers want a second option, particularly after any incumbent stumble.

What could go wrong

Smaller scale than the leaders in a market where scale improves detection.

Position in the ecosystem
publicGOOGL

Wiz (Google)

What it does

Cloud security posture and workload protection, agreed for acquisition by Alphabet.

Why it matters

Cloud misconfiguration remains one of the most common routes into an environment, and AI workloads add new data paths worth watching.

What could go wrong

Integration risk inside a much larger owner, and customers who prefer their security vendor to be cloud neutral.

Position in the ecosystem

Network, access and identity

Controlling who and what may reach which system. Agents need credentials too, which makes machine identity a growing part of the problem.

publicPANW

Palo Alto Networks

What it does

Sells network, cloud and security operations products as an integrated platform.

Why it matters

Security buyers are actively reducing vendor counts, which favours suppliers who can credibly cover several categories at once.

What could go wrong

Platform deals can depress near term billings, and best of breed rivals keep winning individual categories.

Position in the ecosystem
publicZS

Zscaler

What it does

Provides cloud delivered zero trust access, inspecting traffic between users, applications and the internet.

Why it matters

When applications and staff are both distributed, the network perimeter is a poor control point. Inspection in the cloud is the practical alternative.

What could go wrong

Competing against both network incumbents and cloud providers bundling similar controls.

Position in the ecosystem
publicNET

Cloudflare

What it does

Operates a global network providing content delivery, DDoS protection, zero trust services and edge compute.

Why it matters

It sits in front of a large share of web traffic, which gives it both a security vantage point and a natural place to put inference close to users.

What could go wrong

Monetising a wide free tier is a long game, and the developer platform competes with the hyperscalers.

Position in the ecosystem
publicOKTA

Okta

What it does

Provides identity and access management for workforce and customer applications.

Why it matters

Autonomous agents need credentials, and machine identity is becoming as significant a control problem as human identity.

What could go wrong

Identity is bundled aggressively by the large cloud providers, and past security incidents left a mark.

Position in the ecosystem

How to think about the economics

  • GrowthHigh

    How quickly demand in this part of the ecosystem is expanding.

  • Capital intensityLow

    How much money has to be spent up front before revenue arrives.

  • Competitive moatHigh

    How difficult it is for a credible new entrant to take the business.

  • Customer concentrationLow

    How much revenue depends on a small number of buyers.

  • Disruption riskModerate

    How exposed the layer is to a technical or commercial shift.

This is a framework for thinking about the economics of a layer, not a recommendation. An important AI company, a strategically advantaged company, an investable security and an attractively valued security are four different things.

Where value may accrue

Telemetry scale, consolidation preference among buyers, and the operational cost of switching a control that touches everything.

Related questions

What is agentic AI?

Most AI use is still reactive: you ask, it answers. An agent is given a goal instead, and works out the intermediate steps, calling tools, reading data and looping until it decides the task is finished. That shift is why permissions, audit trails and identity suddenly matter so much. A chatbot that is wrong wastes your time. An agent that is wrong can take an action on your behalf, which is a different category of problem.

What is a foundation model?

A large, general purpose model trained on a very broad dataset, which can then be adapted to particular tasks. The GPT, Claude, Gemini and Llama families are all foundation models. They are called foundations because applications, tools and agents get built on top of them. Training one is an expensive bet that sitting at the base of the stack is where lasting value accrues, which is precisely the question the rest of this site is trying to help you think about.

Sources and further reading (2)+
  1. 1Investor relations and threat reporting · CrowdStrike
  2. 2Radar internet traffic and security data · Cloudflare
Last fact-checked: 29 August 2026

The AI ecosystem changes rapidly. Company positions, technologies and market data reflect information available at the date above.