Splunk Platform: Building the Data Foundation for Agentic AI

Hello from Amsterdam! As we kick off Cisco Live EMEA 2026, the energy here is electric. If 2024 was the year of the "Chatbot," 2026 is officially the year of the Agentic AI Revolution.

At Splunk, we’ve always been about helping you find the "needle in the haystack." But as data volumes explode and threats become more sophisticated, just finding the needle isn't enough. You need a platform that doesn't just show you data but acts on it.

Today, we are thrilled to share how the Splunk Platform is becoming the ultimate data fabric backbone for the AI era. From hosted models to the Model Context Protocol (MCP), we’re making sure your data isn't just sitting there; it's working for you.

Building on the momentum from .conf25, we are launching innovations that transform how teams interact with machine data and deploy AI at scale. These advances empower you to build autonomous workflows across SecOps, ITOps, and NetOps.

The Agentic AI Era Demands a Different Data Platform

Here's the reality: AI assistants and agents are proliferating across enterprises at breakneck speed. Gartner predicts that 33% of enterprise applications will incorporate agentic AI by 2028, and IDC reports that 80% of organizations are already investing in it. But there's a catch; these AI systems are only as good as the data they can access.

Native AI Models for Machine Data: Splunk Hosted Models (GA)

Large Language Models (LLMs) are good for general purpose tasks but struggle with understanding the complexity of machine data in your operational environment. A generative AI model that understands machine data requires special model training and architecture, resulting in bigger benefits such as greater accuracy, ready-to-use without further tuning, and flexibility to integrate deeper into your security and observability workflows.

That’s why we are officially launching Splunk hosted AI models (General Availability on February 18, 2026).

By bringing these generative AI models directly into the Splunk Cloud Platform, we’re removing the friction of third-party integrations. You get:

• Foundation-sec-1.1-8b-instruct: A model built specifically for security pros. It understands your logs, helps summarize events, and suggests detections in natural language.
• GPT OSS (20B and 120B): Powerful open-source models for general data exploration and complex reasoning.

Enabling AgenticOps with the Splunk MCP server (GA)

One of the biggest hurdles in AI today is the "Island Problem"—AI tools that can’t talk to your data. Enter the Model Context Protocol (MCP).

On February 4, 2026, we are GA with the Splunk MCP Server. This is a game-changer. It provides a secure, standardized bridge between your intelligent agents and your Splunk data. Instead of building custom integrations for months, you can now connect AI agents to Splunk’s search and discovery tools in minutes.

A Modern Foundation: Machine Data Lake & Latest Splunk Releases

AI is only as good as the data feeding it. To support the "Agentic Era," we’re announcing that the machine data lake will be Alpha in February. This is a schema-less, scalable way to land and enrich telemetry before it’s even indexed, making it "AI-ready" for RAG (Retrieval-Augmented Generation) pipelines and fine-tuning.

The Power of Cisco + Splunk

Everything we’re announcing this week at Cisco Live EMEA is part of a bigger picture: the Cisco Data Fabric architecture, underpinned by the Splunk Platform. By bringing together network, infrastructure, and application data, we are helping you turn raw machine data into a strategic asset in the Agentic era.

And now, whether you are running Splunk Cloud Platform 10.2 or Splunk Enterprise 10.2 (both GA as of January 15!), you now have access to SPL2, next-gen data management UI, and the ability to ingest historical S3 data on-demand via Promote. Talk about a lineup of some super impactful capabilities.

The Bottom Line

The shift from "Chatbots" to "Agentic Systems" is here. Splunk is providing the specialized models, the secure protocols, and the massive data scale needed to lead this charge. We’re not just building a data platform; we’re building the foundation for digital resilience in the AI age.

Stay tuned for more updates throughout Cisco Live EMEA, and if you’re in Amsterdam, come see these innovations in action! Also, feel free to check out more about the Cisco Data Fabric powered by the Splunk Platform, Splunk MCP Server, Splunk AI Toolkit, and the Machine Data Lake.

^{Note: Features described as "alpha" or "beta" are subject to change and provided for informational purposes. Development, release, and timing remain at Splunk's discretion.}