AI Threat Assessment · 26 May 2026

Real-Time Innovations (RTI)

Industrial Middleware
STILL BREATHING
2.8/ 10

“RTI built the invisible nervous system for everything from F-35 fighter jets to Mars rovers — the real-time data plumbing that keeps distributed systems from becoming very expensive paperweights. Twenty-five years of aerospace contracts and safety certifications later, they've become the middleware equivalent of a Swiss bank: boring, essential, and charging accordingly. The problem is that their entire value proposition — 'we make disparate systems talk to each other reliably' — just became a prompt away from obsolete.”

Business Model
4.0
Automation Risk
3.5
Moat Strength
2.0
Adaptability
3.0
Need Survival
2.5
AI Threat Level
BUSINESS MODEL REPLACEABILITY

RTI charges enterprise premiums for Connext middleware that orchestrates real-time data between industrial systems — which sounds bulletproof until you realize that Claude and specialized AI agents can now generate integration code, handle error recovery, and manage data flows without the $100K+ annual licensing fees. The safety-critical certifications buy time, but the core 'system integration complexity' that justifies the pricing is becoming a solved problem.

4.0
WORKFORCE AUTOMATION RISK

Systems engineers who spend months designing RTI integration architectures are already being assisted by AI tools that understand distributed systems patterns — GitHub Copilot can draft DDS configurations, and specialized robotics AI can optimize data flows. The human expertise shifts from 'how to configure middleware' to 'what the robots should actually do,' which RTI doesn't control.

3.5
MOAT STRENGTH

Here's the real moat: DO-178C avionics certification, ISO 26262 automotive safety approval, and 25 years of compliance documentation that would take competitors a decade to replicate. When your middleware failure kills people or crashes billion-dollar systems, 'just use the AI version' isn't an acceptable risk. The regulatory approval pipeline is RTI's Swiss Guard — slow, expensive, and irreplaceable.

2.0
AI ADAPTABILITY SIGNALS

They've pivoted hard to 'Physical AI Systems' branding and launched AI-specific development tools, but the underlying product is still the same DDS middleware with machine learning APIs bolted on. The real test isn't whether they can integrate with AI — it's whether AI-native distributed computing frameworks make their entire architecture unnecessary.

3.0
WILL THE NEED SURVIVE AI?

Real-time data coordination between physical systems survives — F-35s will always need their sensors talking to weapons systems in microseconds, and Mars rovers need fault-tolerant communication between subsystems. The question is whether future autonomous systems will use RTI's middleware architecture or AI-designed protocols that make traditional pub-sub messaging look like carrier pigeons.

2.5
Verdict

“The regulatory approval moat protects RTI for now, but every new autonomous system designed with AI-native architecture is a customer they'll never win back — the transition happens one procurement cycle at a time. They're not building the nervous system for the age of autonomous everything; they're the last great middleware company before AI designs better middleware than humans ever could.”

Scores are based on public information and AI analysis. This is an affectionate roast, not a financial assessment. The best companies use this as a mirror, not a verdict.

Roast another →