When Tools Become Products: How GhostGrid's Internal Software Became a Business

When we announced GhostGrid -- our universal free-space optical networking platform with 20 provisional patents -- the response was immediate. Investors asked about hardware timelines. Engineers asked about simulation data. But the question nobody asked turned out to be the most important one: what happened to all the software you built to get there?

The answer: it became its own business.

The Problem Nobody Plans For

Building a hardware-focused R&D project like GhostGrid requires simulation. Before you bend metal or grind glass, you validate designs computationally. So we built tools -- a beam propagation simulator here, an atmospheric turbulence model there, a mesh network topology analyzer somewhere else.

By the time we'd validated our 20 patent-protected innovations, we had 53 separate tools scattered across scripts, notebooks, and containers. That's when we realized something: we hadn't just built validation infrastructure. We'd built a physics simulation platform.

The same realization hit twice more -- once when our patent portfolio grew beyond what spreadsheets could manage, and again when USPTO figure requirements demanded precision we couldn't achieve with general-purpose drawing tools.

Each time, the pattern was the same: build an internal tool to solve an R&D problem, discover the tool solves a broader market problem, productize it. This is the story of how that happened three times in six months.

Act 1: The Optikal Physics Suite

From 53 Scripts to 74 Professional Tools

GhostGrid's original simulation toolkit -- the "GhostGrid Physics Suite" -- started as a collection of Python scripts for validating specific innovations. The VR Lighthouse Alignment System needed beam steering simulation. The Hybrid Cognitive FSO-RF Network needed atmospheric modeling. Each patent required its own validation pipeline.

The scripts worked. But they were fragile, poorly documented, and impossible for anyone outside the project to use. When we stepped back and looked at what we'd built, we saw coverage across optical physics, orbital mechanics, network topology, 3D spatial processing, machine learning, and more.

In November 2025, we rebranded it as the Optikal Physics Suite and made a decision: treat it as a product, not a side effect.

We containerized the physics engines. We standardized interfaces. We added five entirely new scientific domains -- electromagnetic simulation, computational fluid dynamics, robotics, finite element analysis, and advanced optimization. The tool count grew from 53 to 74 tools across 15 scientific domains.

What It Is Today

The Optikal Physics Suite packages 74 tools into a unified platform:

Category	What's Included
Core Scientific Computing	NumPy, SciPy, Pandas, Numba for high-performance numerical work
Optical Physics & FSO	Beam propagation, atmospheric effects, link budget analysis
Orbital Mechanics	Satellite constellation simulation, orbital debris tracking
Network Topology	Graph theory, mesh optimization, routing simulation
3D Spatial Processing	LiDAR point cloud analysis, Gaussian Splatting
Machine Learning	Reinforcement learning for adaptive optics and routing
Electromagnetic Simulation	Antenna modeling, RF propagation (via GNU Radio, OpenEMS)
CFD & Weather Modeling	Atmospheric prediction for proactive FSO failover (via WRF, OpenFOAM)
Finite Element Analysis	Structural simulation for enclosure and mount design (via FEniCSx)
Robotics & Control	Autonomous alignment systems (via ROS2, Drake)

Seven containerized physics engines handle the heavy computation: Drake, FEniCSx, GNU Radio, OpenEMS, OpenFOAM, ROS2, and WRF. The entire suite runs 1,892 tests at 97% pass rate with 81% code coverage.

All 74 tools are free and open-source under the MIT license. We position it as "MATLAB + STK + COMSOL for the cloud generation" -- the same simulation capabilities that cost six-figure annual licenses, available at no cost to researchers and engineers.

The Business Model

Free tools generate revenue in three ways:

1. Enterprise support and managed deployment -- Organizations that need guaranteed uptime, custom configurations, and priority support pay for service tiers.
2. Consulting engagements -- Companies building their own optical or physics-heavy systems hire us to configure and extend the suite for their specific problems.
3. Credibility and deal flow -- When a prospect evaluates our services, a 74-tool simulation platform demonstrates technical depth that no pitch deck can match.

The Optikal Physics Suite was never on our product roadmap. It exists because GhostGrid needed simulation infrastructure. Now it generates revenue independently of GhostGrid's hardware timeline.

Act 2: PATLAS and the Patent Management Problem

When 20 Patents Need a System

Our GhostGrid announcement described 20 provisional patent applications. Twenty patents sounds manageable -- until you account for filing deadlines, prior art searches, figure requirements for every claim, relationships between innovations, and portfolio strategy decisions across multiple technology domains.

At scale, even 20 patents generate hundreds of administrative touchpoints. And as additional innovations emerge from continued R&D and simulation work, the complexity grows faster than headcount.

We needed a patent management system. Every commercial option we evaluated was either built for law firms (wrong workflow), prohibitively expensive (wrong economics for a startup), or missing the technical integration we needed (can't connect to our simulation data).

So we built PATLAS.

Patent Atlas: Portfolio Management for Technical Teams

PATLAS (Patent Atlas) is a full-lifecycle patent portfolio manager built for technical teams, not law firms. It tracks every patent from provisional filing through prosecution, manages relationships between innovations, and orchestrates the documentation pipeline.

The architecture reflects our stack: TypeScript monorepo with Turborepo, GraphQL API with real-time WebSocket subscriptions, PostgreSQL with Drizzle ORM, and Redis caching. It's the same enterprise-grade infrastructure we use for everything else at Netrun Systems.

Key capabilities:

• Portfolio dashboard -- Visual overview of all patents by status, domain, filing date, and strategic priority
• Deadline management -- Automated tracking of provisional-to-nonprovisional conversion windows, response deadlines, and maintenance fees
• Innovation linkage -- Maps relationships between patents, showing how one innovation builds on and reinforces others across the GhostGrid platform
• Prior art integration -- Connected to KAMERA (our AI research platform) for automated prior art monitoring across 100M+ documents
• Figure orchestration -- Manages the technical drawings required for every patent application (more on this below)

KAMERA: AI-Powered Prior Art Research

Before PATLAS, prior art searches were manual. An engineer would spend hours on Google Patents and USPTO databases, trying to determine whether a particular GhostGrid innovation was truly novel.

KAMERA automates this process. Originally built as Netrun's AI-powered job search and research tool, we extended it for intellectual property -- scanning patent databases, academic papers, and technical publications to identify potential prior art before we invest in a full patent application.

When a GhostGrid simulation reveals a potentially patentable approach, KAMERA runs automated prior art analysis. If the innovation appears novel, PATLAS creates a tracking entry and begins the documentation workflow. If prior art exists, we know before spending money on attorneys.

Act 3: FIGR and the Figure Drawing Problem

USPTO Doesn't Accept Napkin Sketches

Patent applications require technical figures -- precise, labeled diagrams that illustrate every claim. The USPTO has specific requirements for line weights, reference numerals, shading patterns, and drawing standards.

For a handful of patents, outsourcing figure generation to a patent illustrator is feasible. For 20 patents -- each requiring multiple detailed figures across beam steering systems, mesh topologies, wavelength management architectures, and orbital mechanics -- the cost and turnaround time become prohibitive. And as the portfolio grows, the problem only compounds.

We needed a tool that could generate USPTO-compliant figures programmatically from our technical data. Nothing commercially available connected to our simulation outputs or understood optical network diagrams.

FIGR: Programmatic Patent Figures

FIGR (Figure Renderer) is a React/SVG rendering engine purpose-built for patent illustration. It takes structured data -- from PATLAS patent records, Optikal simulation outputs, or manual specifications -- and produces publication-ready SVG diagrams that meet USPTO drawing standards.

The tool handles:

• Reference numeral management -- Automatically assigns, tracks, and consistently places reference numerals across multi-figure patent applications
• Standard compliance -- Line weights, margins, shading patterns, and labeling that conform to USPTO Manual of Patent Examining Procedure requirements
• Technical accuracy -- Figures generated from actual simulation data, not artistic interpretations. When a figure shows beam propagation through atmospheric turbulence, the visualization reflects real physics
• Batch generation -- Produce all figures for a patent application in a single pass, with consistent styling and cross-referencing

FIGR eliminated a bottleneck. What previously required weeks of back-and-forth with illustrators now takes hours of engineer time. The figures are more accurate because they're derived from simulation data, and they're more consistent because the rendering engine enforces standards programmatically.

The Pattern: R&D Infrastructure Becomes Product

Three tools. Three products. One pattern.

R&D Need → Internal Tool → Market Gap Discovery → Product

Internal Need	Tool Built	Market Problem Solved
Validate GhostGrid physics	Optikal Physics Suite	Researchers lack affordable simulation platforms
Manage growing patent portfolio	PATLAS + KAMERA	Technical teams lack IP management tools
Generate USPTO-compliant figures	FIGR	Patent applicants lack programmatic figure tools

This isn't accidental. It's a structural advantage of building deep technology.

When you're solving hard problems -- real physics simulation, complex portfolio management, regulatory-compliant document generation -- the tools you build to support your process are themselves solutions to problems that other organizations face. The R&D investment doesn't just produce the primary innovation (GhostGrid). It produces the supporting infrastructure, and that infrastructure has independent value.

How Tools Fund Labs

Here's the economics that matter for a bootstrapped R&D company:

GhostGrid requires hardware prototyping -- an expensive, time-intensive next step that depends on seed investment. The timeline is measured in years.

Optikal Physics Suite, PATLAS, and FIGR are software products that can generate revenue today. They require no hardware, no manufacturing, and no physical distribution. They're deployable immediately because they're already deployed -- we use them daily.

Every consulting engagement sold through the Optikal Physics Suite funds further GhostGrid simulation work. Every PATLAS subscription helps cover patent filing costs. Every FIGR deployment reduces our per-patent illustration expenses.

The tools fund the lab. The lab produces more tools. The cycle accelerates.

What This Means for GhostGrid

GhostGrid started Phase 1 with 20 validated innovations and a 53-tool simulation suite. Six months later:

• 20 provisional patents filed with a pipeline of additional innovations emerging from deeper simulation
• 74 simulation tools (up from 53) -- five new scientific domains added
• 3 independent software products launched from R&D infrastructure
• Revenue-generating tools that reduce GhostGrid's dependency on external funding

The hardware roadmap hasn't changed. We still need seed investment for prototype development. But the path to that prototype is now supported by software products that didn't exist six months ago -- products that generate revenue, demonstrate technical capability, and validate our engineering approach.

Try It Yourself

• Optikal Physics Suite: Open-source, MIT licensed. Available now for researchers, engineers, and technical teams building physics-intensive systems.
• PATLAS: Patent portfolio management for technical organizations. Contact us for early access.
• FIGR: Programmatic USPTO-compliant patent figure generation. Available as part of the Netrun Patent Suite.

If you're building something that requires simulation, IP management, or regulatory-compliant documentation -- or if you're interested in how GhostGrid's R&D pipeline translates to investment opportunity -- visit netrunsystems.com or reach out directly.

---

Netrun Systems is a California C Corp building infrastructure for the next era of connectivity. GhostGrid is our hardware R&D program with 20 provisional patent applications filed with the USPTO. Optikal Physics Suite, PATLAS, and FIGR are production software products available today. Tool specifications reflect current status as of February 2026.