Cisco acquires NeuralFabric to expand AI capabilities for businesses. The acquisition should help organizations build, train, and deploy specific AI models and Small Language Models (SLMs) within their own infrastructure. NeuralFabric’s technology should primarily ensure that the new AI Canvas has an even more solid foundation.
The future of AI models lies at least as much in small models as in large ones. To make AI truly interesting within organizations, we don’t need another generic model, but rather more specialized models and SLMs. That’s what Cisco is fully committed to, at least.
We have already seen examples of this with the Security Reasoning Model and the Deep Network Model. The first model focuses specifically on security issues and is based on more than 30 years of telemetry from Cisco itself. The second model is an important pillar of AgenticOps developments at Cisco, specifically AI Canvas. We recently wrote an extensive article about the latter.
NeuralFabric is a new piece of the agentic puzzle
The theory of specialized models and SLMs certainly sounds good. However, it also raises some questions. When you talk about specifically trained models, this usually means that they are also intended for specific environments. And by specific environments, we mean the environments of organizations themselves. The models that need to be trained for this also use the organizations’ own data.
NeuralFabric, a company based near Seattle and founded in 2023, has developed a platform that enables organizations to do this. That makes it an interesting addition to the agentic puzzle that Cisco is working on.
Specifically, AI Canvas, which already uses the Deep Network Model, can benefit greatly from this. After all, organizations can also use this in their own environment for purposes other than troubleshooting their own network. However, this is only possible if there is a good and, above all, reliable and compliant foundation underneath.
NeuralFabric as the foundation for AI Canvas
NeuralFabric is set to become this foundation for AI Canvas. It promises complete control and transparency when it comes to the data that organizations use. It is possible to combine public datasets with private datasets and synthetic data. NeuralFabric also guarantees that the organization’s own data will never be disclosed. It also claims to have a modular offering, making it possible to add AI components one by one. Finally, according to NeuralFabric, it should be 90-99 percent cheaper than deploying one of the generic, large models, and it is possible to build an SLM in a matter of hours.
All in all, the integration of NeuralFabric into Cisco’s offering should enable organizations to build faster, better (and compliant) models for their own environments than is currently the case. Developments in LLMs are far from complete, but SLMs are undoubtedly becoming increasingly important. It is therefore a good idea to respond to this trend. Cisco is doing so with this proposed acquisition. It is expected to be completed in Q2 of Cisco’s fiscal year, i.e. before the end of January 2026. No details have been disclosed about how much Cisco is paying for NeuralFabric.