NetFabric - AI Agents for Network Operations

TL;DR

TL;DR: NetFabric is an ETH Zurich spin-off that deploys autonomous AI agents for network operations — the agents investigate incidents, coordinate across data sources (logs, configs, DBs), and can take action to resolve issues automatically.

Source and Accuracy Notes

⚠️ This section is MANDATORY. All links must be verified from actual source, not guessed.

• Project page: netfabric.ai — verified via curl
• HN launch thread: news.ycombinator.com/item?id=41068183 — 6 points, “next-gen network monitoring” by pentab
• License: proprietary (no open-source version available as of July 2026)
• Funding: CHF 1.2M Innosuisse Innovation Grant (April 2026); $2.2M pre-seed (July 2024) — verified from netfabric.ai/news
• ETH affiliation: official ETH AI Center affiliate, ETH Zurich Spin-Off label (July 2024)

What Is NetFabric?

NetFabric is a network operations platform built around autonomous AI agents that move beyond passive monitoring. Where traditional tools alert on symptoms, NetFabric agents investigate root causes by querying multiple data sources in parallel, then coordinate next steps — sometimes acting automatically to resolve issues.

The product is built by an ETH Zurich spin-off with a PhD-heavy team (CEO recognized for outstanding dissertation in Programming Languages, July 2024). The company has backing from Innosuisse, Switzerland’s national innovation agency.

How It Works

NetFabric agents operate across three layers:

Agent Layer

The AI agent itself, which reasons about incidents and orchestrates the investigation plan. It decides which modules to call, in what order, based on the context of the incident.

Modules

Domain-specific agents specialized for different data sources:

• Database Module — queries network state from operational databases (9.8k rows reduced to 312 via intelligent summarization)
• Config Module — reads and compares device configurations across the network
• CLI Module — executes commands on network devices and parses structured output
• System Logs Module — ingests and queries system logs for correlation

Data Sources

NetFabric connects to:

• Network device CLIs (SSH-based command execution)
• Configuration databases
• System and syslog sources
• Any structured data store the network team uses

Key Features

Live Demo Animation: The site features an animated terminal that visualizes a complete incident investigation cycle — from detecting a connection instability to identifying optical degradation between specific network nodes (AMS-3.1 and PAR-10.2), completing in under 2 minutes.

Dual Deployment: The agent runs either as a cloud-hosted service or can be deployed on-premises, fitting different security postures.

Multi-vendor Support: Works across network hardware vendors rather than being tied to a specific OEM.

Scale Without Headcount: NetFabric’s core pitch is that one agent can handle investigations that would otherwise require a senior NOC engineer — scaling operations without scaling the team.

Practical Evaluation Checklist

• [ ] Connects to existing network monitoring stack (no rip-and-replace)
• [ ] Investigates incidents autonomously within defined safety guardrails
• [ ] Supports multi-vendor network hardware
• [ ] Provides on-premises deployment option for air-gapped environments
• [ ] Integrates with existing ticketing and alerting pipelines
• [ ] Reduces noise vs. traditional threshold-based alerting

FAQ

Q: Is NetFabric open source? A: No. As of July 2026, NetFabric is a proprietary platform with no public open-source version. The company is focused on the cloud and on-premises product.

Q: What networks does it support? A: NetFabric works with network device CLIs via SSH, configuration databases, and syslog sources. It is vendor-agnostic at the protocol layer.

Q: How is it different from standard network monitoring tools? A: Traditional monitoring alerts on symptoms (high latency, packet loss). NetFabric agents actively investigate root cause by querying multiple data sources simultaneously, building a coherent picture of what went wrong and what the fix sequence is.

Q: Is there a free tier or self-hosted option? A: The site promotes “Book a live demo” as the onboarding path. No self-service pricing page is available publicly as of July 2026.

Q: Where is the company based? A: NetFabric is based in Switzerland, is an official ETH Zurich Spin-Off, and is affiliated with the ETH AI Center.

Conclusion

NetFabric represents a shift in how network operations is done — from human-driven investigation to AI-coordinated automation. With strong academic backing (ETH Zurich), public funding (Innosuisse), and a clear product demo, it is one of the more credible players in the AI-for-NetOps space. The “book a demo” gating makes it hard to evaluate hands-on without a sales conversation, but the animated product demo on the homepage is worth watching even just to understand the investigation flow.

For teams running large network infrastructures who are tired of alert fatigue and slow root-cause analysis, NetFabric is worth a conversation.