Enterprises are all-in on AI, but most are still trying to figure out how to do it responsibly, efficiently, and without running into barriers like data privacy rules or limited access to GPUs.
Cisco acquires NeuralFabric to extend AI model development opportunities
Cisco’s latest move is aimed directly at those challenges. The company intends to acquire NeuralFabric, a Seattle-area startup focused on helping companies build their own generative AI models using their own data.
NeuralFabric was founded in 2023 by former Microsoft veterans and has built technology that “enables organizations to develop domain-specific Small Language Models (SLMs) using their own proprietary data—deployable across SaaS and on-premise environments with the flexibility and control that today’s enterprises demand,” according to DJ Sampath, Cisco’s senior vice president of AI software and platforms.
Goodbye, one-size-fits-all AI
Cisco says we’re well past the stage where a single massive model trained on the entire internet can meet every enterprise need. Companies want AI that understands their world, which means their processes, industry rules, and compliance requirements.
That idea already lives inside Cisco’s AI Canvas, which the company describes as a collaborative workspace for teams and AI agents. AI Canvas works with Cisco AI Assistant and domain-aware models, and NeuralFabric’s technology slots neatly into that vision by making it easier to build and deploy specialized models right where customers need them.
As Sampath put it, the startup has “cracked a crucial part of this puzzle” by giving enterprises more control and flexibility in how they “operationalize” AI.
Strengthening a growing AI strategy
Cisco has been very clear that it is placing big bets on AI, including a $1 billion investment fund to back promising startups.
NeuralFabric brings two significant advantages to that momentum:
- A modular platform for developing Small Language Models (SLMs)
- Expertise in distributed systems, model training, and flexible deployment
Cisco says these capabilities will “help develop and strengthen the foundation of AI Canvas” by improving model customization, speeding up training, and supporting hybrid deployment environments that meet data sovereignty needs.
Looking ahead
Once the deal closes, NeuralFabric’s team will join Cisco’s AI Software and Platform organization. Both companies will continue operating independently for now, but the goal is aligned: to make enterprise AI practical, secure, and smarter about the environments it serves.
Cisco summed it up this way: “The development of SLMs and domain-specific AI models like the Deep Network Model are just the beginning of Cisco’s commitment to our customers in the AI era.”
The NeuralFabric deal lines up with Cisco’s recent push into AI-ready infrastructure. Those platform upgrades focused on helping customers run AI at scale; this acquisition focuses on helping them build AI that fits their business. Taken together, Cisco is positioning itself to support both sides of the enterprise AI equation.