Promise made, promise kept! At last year’s Splunk Conference, Cisco executive Jeetu Patel proclaimed boldly: “We’re not going to screw up Splunk!” Wow! What a thing to say!

As a career journalist, I had no choice but to focus on this remarkable vow, and for several reasons. First, it’s a bold thing to state! Second, it shows that he knew there was trepidation amongst the Splunkers, who must have been pondering their fate.

And third? Because it showed a remarkable awareness by a major corporate executive that his company had, at times, found itself in the crosshairs of criticism for acquiring companies, but failing to follow through on integrating their tech, and teams, in a meaningful way.

Anyone remember Composite Software? Yeah.

But by all accounts, Patel has lived up to his promise! At this year’s .conf (their clever name for the annual get-together), he came right out and said that Splunk is a strategic asset for Cisco, and will be going forward. And that makes total sense, to the tune of $28 billion!

In many ways, the Cisco-Splunk union is a match made in enterprise Heaven! After all, Cisco builds the routers and switches that power businesses the world over; while Splunk captures and processes the data necessary to keep all that machinery running efficiently.

Yes, Splunk is a security company. But really, they’re an observability firm that facilitates security, governance, and compliance, while also helping with that all-important aspect of enterprise technology: performance!

“Think about that rage-clicking moment,” mused Mimi Shalash, Splunk’s manufacturing maven. In a panel discussion near the end of the show, she pointed out that some of her clients have saved $50 million thanks to the efficiency gains they’ve achieved with Splunk.

The thing about performance is that it always matters, whether for security, processing, governance, whatever. By giving visibility into the complex systems that run the modern enterprise, Splunk enables better performance across the board. That’s a really big deal.

Convergence

Veteran journalist Mike Vizard, now serving as Chief Content Officer for Techstrong Group, asked the panel about the blurring of lines between traditionally discrete disciplines, ike monitoring, analytics, observability and security.

Splunk’s platform has always excelled at stitching together telemetry from diverse sources. Now, with Cisco’s network reach and AI advances, that ability is expanding. Patel described the challenge like this:

“It’s never clear what exactly is going on when an incident occurs,” he said Often, IT teams juggle separate tools, data, and silos. The Splunk-Cisco approach brings them together, creating a “consistent, cohesive, correlated data guide model” that spans infrastructure, applications, and security.

Patel also noted the rise of AI agents as a force that will reshape how applications are designed and secured. This next generation of apps are “not going to be optimized just for humans. They’ll be optimized for humans and agents.”

That shift requires new forms of identity and permissioning, where fine-grained controls are applied not just to people, but to AI systems acting on their behalf. “If we can do that elegantly, we’ll redefine how AI delivers value to the enterprise,” said Patel.

New Foundations

Splunk used the Boston .conf25 stage to introduce something unprecedented: a Time-Series Foundation Model for machine data. Unlike general-purpose LLMs, this model is trained specifically on the telemetry that runs modern enterprises: logs, metrics, events, and traces.

The goal is to give security and operations teams an AI-native way to detect anomalies, forecast issues, and automate responses at machine speed. By focusing on time-series and event-driven data rather than natural language, Splunk is staking a claim that AI for enterprise resilience requires its own kind of foundation model.

Cisco emphasized openness as well: the new model will be made available through Hugging Face, encouraging community adoption and customization. That decision signals their intent to contribute to a broader ecosystem of AI-ready tooling for IT, security, and observability.

As enterprises race to build “agentic” AI capabilities, this foundation model provides a common baseline that can be tuned for use cases like anomaly detection, automated forecasting, or threat hunting. In short, it’s a major step toward embedding machine intelligence directly into the telemetry layer of the modern digital business.

The model was trained on multiple streams of telemetry: sensors, applications, networks, and security logs. Patel stressed the importance of making AI “machine data ready”, pointing out that machine data represents 55% of data growth yet remains underutilized in AI.

Not everyone is optimistic about the potential here, however. One prominent industry analyst noted: “Nobody told Splunk that deep learning doesn’t work well on time series or tabular data. The papers were published back in the early 2000s.”

But hey, hope springs eternal! “Machine data is the fuel for AI, and Splunk sits right at the heart of that opportunity,” said Patel.

On Brand

Perhaps most importantly from a pure Splunk perspective, there were many open commitments to supporting not just the technology, but the brand itself. Said Patel: “Splunk has so much brand equity, there’s no reason to roll them directly into Cisco.”

“Usually when a giant buys a smaller company, innovation slows. With Cisco and Splunk, the opposite has happened. We’re supercharging innovation. The combination of Cisco and Splunk brings together all six critical success factors: timing, market, team, product, brand, and distribution. This is the most exciting time in Cisco’s 40-year history.”

He concluded: “The Splunk community is one of the strongest in tech, and we are deeply committed to preserving and growing it. We’re just getting warmed up.”