Cisco Live: What matters most in startups

In Silicon Valley, venture capital (VC) once meant scouring garages and coffee shops for the next whiz kid with a napkin sketch of world domination. The pitch comes with polished decks and a pre-trained language model. If it can’t explain its AI advantage in 30 seconds, it rarely gets a second look.

The senior vice president of corporate development at networking giant Cisco, Derek Idemoto, who also heads up the Cisco Investments VC arm, has watched this shift from the front row. His portfolio spans more than 250 active investments.

With acquisitions and investments totalling over $50-billion, he has helped shape Cisco’s approach to innovation for nearly two decades. His fingerprints are on some of Cisco’s most significant strategic moves. Most recently, he led Cisco’s $28-billion acquisition of cybersecurity and observability giant Splunk – the company’s biggest buyout yet.

Rather than a quest to find the next Google, Idemoto is more about aligning emerging talent with Cisco’s platform and product strategy. That includes AI, which now consumes most of his deal flow.

“One year in AI feels like a lifetime,” he tells me over coffee (and a blank napkin)  at the Cisco Live 2025 conference in San Diego this week. “We launched the $1-billion Global AI Investment fund last year to move fast, scale responsibly, and build lasting partnerships. We’ve already backed a couple dozen companies, and that momentum continues.”

This is corporate venture capital with a mission. Cisco picks strategic collaborators, and focuses on those that build real infrastructure and solutions. For example, it joined an investment round to pump $640-million into AI hardware startup Groq – no relation to Elon Musk’s AI chatbot, Grok – to give it a $2.8-billion value.

Two-thirds of new investments now follow an AI-native path. The other third might focus on cybersecurity, observability, or new approaches to data flow, often with AI woven into the core.

“If a company brings a clear AI story, it enters a different tier of interest. That story doesn’t need to rely on hype. It needs to connect with enterprise priorities.”

He points to a current opportunity in next-gen cybersecurity.

“There’s a startup approaching risk visibility in a way that feels fresh. That stands apart from investments like Anthropic, which tackle foundation model design. Each company addresses different layers of the AI stack.”

Investments often grow into acquisitions.

“When we invest and later acquire, we’ve already seen how the team operates, how the culture fits, and how the product scales. That allows us to integrate with confidence.”

Even when the path leads to an IPO instead of M&A, the benefits persist. Core weave, an AI infrastructure company that has seen its share price more than treble since its March listing, is a case in point. Last year Cisco Investments joined a funding round that valued the company at $28-billion.

“We backed CoreWeave before their IPO. Their valuation multiplied, and our partnership deepened. Their CEO, Mike Intrator, remains a close friend of Cisco. Those kinds of relationships don’t fade.

 “We invest to learn, to build trust, and to deepen our understanding of future use cases. The companies we work with bring energy, speed, and perspectives that challenge our assumptions.”

He speaks with calm precision, in the voice of a strategist who sees market shifts as patterns rather than surprises.

“I’ve met over 100,000 startups across my career. That helps me recognise what sets great founders apart: clarity of purpose, strong instincts, and a willingness to adjust course without losing direction.”

His map of the near future includes agentic AI, edge-native systems, and robotics in forms that blend hardware, software, and autonomy.

Looking further ahead, he has a tantalising vision: “I’m excited to see what enabling technologies will accelerate the speeds and feeds of a network, which often are the limiting factor. What can we solve if Quantum comes together with optics, silicon, photonics, things like that?

“If you have unconstrained bandwidth, unconstrained compute, imagine what problems we can solve? What diseases we can we cure? How can we bring people closer together? All that is the benefit once you have unconstrained infrastructure.”

When asked what AI superpower he would choose for himself, he is surprisingly down to earth: “Human connection. I want technology to give me more time with the people I care about.”

* Arthur Goldstuck is founder of World Wide Worx and author of “The Hitchhiker’s Guide to AI”.