PatrolOS

Counter-UAS planning for critical sites.

patrolos.io

02 / 14
Overview

Hostile threats move fast.
PatrolOS moves before.

Unpredictable patrol and response plans that close gaps in counter-drone coverage before an attack can exploit them.

03 / 14
Problem

Critical infrastructure is under attack.

Refinery attack aftermath
2019–2026 / Gulf refineries

The 2019 Abqaiq strike cut half of Saudi oil output overnight. Refineries have been under continuous threat since.

Port oil infrastructure attack
May 2025 / Port Sudan

UAS strikes shut down port operations and supply chains. Offline in hours due to low-cost weapons.

Industrial site strike
Mar 2026 / Mina Al-Ahmadi

Major refinery struck alongside concurrent UAS threats targeting critical AI data-centers.

Adversaries don't overpower the defenses. They slip through the gaps.
04 / 14
Problem · 02 / 02

A fixed defense has a fixed blind spot.

Counter-drone assets are placed once and rarely moved. An adversary only has to watch.

Hardware exists

The sensors are there

Sites already own radars, RF detectors, cameras and response drones. This is the physical layer of a counter-drone system.

Coverage is static

The plan is fixed

Those assets are positioned and scheduled in a way an observer can learn. These patterns repeat on a knowable cycle.

The opening is predictable

The gap is exploitable

A known gap, at a known time, is an invitation. Given enough observation time, any static schedule can be exploited.

Predictability is the vulnerability. The fix is a smarter plan.
05 / 14
Solution

PatrolOS plans coverage that an adversary cannot learn.

01
Take what the site has

Sensors, response drones, teams, terrain, no-fly zones and rules of engagement.

02
Plan adaptive coverage

Assets positioned and scheduled with deliberate variation. No sector predictably open at a knowable time.

03
Re-task during an incursion

When a track appears, PatrolOS assigns response and repositions the rest. No new gap opens. Every decision logged.

It is the planning layer. It does not detect or jam. It decides where your assets go and when they move.
06 / 14
How it works

From site intelligence to patrol missions.

patrolos.io/console
PatrolOS planning console - site map with patrol zones and deployment panel
▶ Watch the demo video
<1s
Plan generation for routine patrols
<15s
Time from threat detected to first defender moving
~100%
Priority zones covered at all times
<7 min
Maximum time any zone is left unprotected
07 / 14
Clients

Customers with critical sites to protect.

Energy - oil refinery
Energy & utilities

Power plants, refineries and pipelines under record attack levels. CIP-014 now mandates demonstrable airspace risk management.

Data centers - server room
Data centers

AI-era hyperscale campuses have high-value, fixed, observable perimeters facing rapidly rising adversarial attention.

Ports - intermodal containers aerial
Ports, airports & defense

Probed and disrupted by drone incursions across Europe since 2025. The budget exists, but the plan does not.

08 / 14
Market

An $80B market where the planning layer is missing.

TAM
$80B → $129B

Global perimeter security, 2025–2030, growing 10.1% per year.

Mordor Intelligence, 2025

SAM
$4.8B / yr

~40,000 high-criticality sites in North America and Europe at ~$120k average annual license.

SOM · 2035
$24M ARR

165 sites. <0.5% of SAM. Recurring licenses of $90k–$250k per site per year.

Management plan

Where the money goes

Hardware dominates

Spend flows to radars, RF sensors, jammers and interceptors. Software planning is a thin line item.

The planning gap

The commercial layer is unserved

Existing C2 platforms are built for government defense programs at government contract prices. Refineries, ports, airports and data centers have no purpose-built planning tool. The gap is commercial and structural.

09 / 14
Business model

Annual software license per protected site.

Paid pilot
~$30–50K
3–6 months · single site
Runs on the customer's existing hardware and data
Annual license
$90–250K / yr
Per site, per year
Recurring revenue · high software margins
Portfolio growth
×N sites
As coverage expands across the portfolio
No new hardware required
Software margins
Recurring revenue
No hardware dependency
Expands with the customer

Exact prices shared in conversation. These are indicative ranges, not commitments.

10 / 14
Competition

The market sells detection and defeat. We sell the plan.

Capability
Detection
Defeat
C2
PatrolOS
Detects incoming drones
✓
✗
✓
✗
Neutralizes active threats
✗
✓
✗
✗
Fuses sensor feeds into one picture
✗
✗
✓
✗
Decides where assets are placed
✗
✗
✗
✓
Adapts patrol plans to threats in real-time
✗
✗
✗
✓
Prevents predictable patrol patterns
✗
✗
✗
✓
11 / 14
Milestones

Three phases over the next 24 months.

Phase 1
Build
Months 0 – 9
Hire a small senior team, ship the planning engine and console, integrate with two sensor types.
Phase 2
Prove
Months 6 – 18
2–3 paid pilots at live sites (at least one airport and one energy or defense site) with measured gap reduction.
Phase 3
Convert
Months 15 – 24
Pilots to annual licenses, reference cases published, Series A raise on hard evidence.
12 / 14
Beyond counter-drone

The same planning engine applies across security verticals.

Civil security - law enforcement drone deployment
Civil security

Patrol allocation for campuses, stadiums and public spaces. The core scheduling problem is identical to counter-drone.

Force protection - deployed armored unit
Defense

Perimeter planning for fixed installations and deployed or moving units. FOBs, convoy corridors, rapidly established perimeters where unpredictability is a tactical requirement.

Forestry - illegal logging deforestation aerial
Forestry & conservation

Illegal logging detection and habitat monitoring across large, varied terrain. Adaptive scheduling with limited autonomous assets and mixed sensor types.

Additional verticals include maritime port security, critical infrastructure perimeter monitoring, and large-scale event security.

13 / 14
Team

Built by adversarial-optimization founders.

Nathalia Wolf
Nathalia Wolf
CEO
  • PhD, Operations Research, Inria
  • MSc, Industrial Engineering
  • Led an Inria Startup Studio project (~€80k secured)
  • 5+ years in Stackelberg security games, patrol allocation and drone autonomy
Juan Sepulveda
Juan Sepulveda
CTO
  • PhD, Computer Science, Inria
  • MSc, Electrical Engineering
  • Built a mixed-integer bilevel optimization solver in C/C++ (branch-and-cut, presolve, heuristics)
  • Postdoc on decentralized multi-agent optimization on Ethereum 2.0

PatrolOS

Counter-UAS planning for critical sites.

patrolos.io

wolf@patrolos.io