Martin Casado

Who they are

Martin Casado is General Partner at a16z — co-founded Nicira Networks, the software-defined networking company VMware acquired for $1.26 billion in 2012, and holds a PhD from Stanford.

Person

Martin Casado joined a16z in 2016 as its ninth General Partner, named directly from VMware where he had served as GM of Networking and Security and SVP and Fellow after the Nicira acquisition. He spent his formative years at Stanford — a Master's in 2005, then a PhD in Computer Science/Networking completed in 2008 — and co-founded two companies while still in academia: Illuminics Systems, an IP analytics company acquired by Quova in 2006, and Nicira Networks in 2007, the SDN company that put software-defined networking on the map and sold to VMware for $1.26 billion in 2012. After the acquisition he stayed at VMware as CTO for Networking and Security before making the jump to VC. The through-line is deep infrastructure bets: networking, then cloud, now AI — each time with a technical founder's conviction about where the stack is shifting. He publishes broadly — TechCrunch, the a16z blog, Forbes, and The Wall Street Journal — and his public voice is sharply policy-focused: in late 2024 he argued in Fortune and TechCrunch that AI regulation should be grounded in evidence, not existential anxiety. On the podcast circuit he's been everywhere from Invest Like the Best to 20VC to The Generalist, consistently pushing the thesis that the AI boom still has years to run — comparing the current moment to 1996 — and interrogating where value will actually accrue in the stack.

Network

Casado's most visible collaborator at a16z is Marc Andreessen — they've co-appeared on podcast discussions about AI's trajectory. He also works closely with Sarah Wang, with whom he co-presented on AI value accrual. Harry Stebbings has engaged publicly with him on AI and competitive-dynamics thinking.

Marc Andreessen· Co-founder and General Partner, a16z
Sarah Wang· General Partner, a16z
Harry Stebbings· Founder, 20VC

How they likely show up

PhD in Computer Science/Networking from Stanford, followed by founding Nicira → brings an engineer's first-principles rigor to investment theses, not just pattern-matching on financials.
Hybrid role type (GP who writes, speaks, and co-presents) → operates as both investor and public intellectual; comfortable being the face of a position.
Prolific across TechCrunch, Forbes, WSJ, and the a16z blog on the same cluster of topics → thinks in frameworks he wants to road-test publicly before they harden into conviction.
Co-founded two companies (Illuminics and Nicira) while still at Stanford → high tolerance for ambiguity, used to building before the market exists.
Consistent 2024–2025 podcast presence hammering a single thesis (AI boom = 1996, value accrual at the infrastructure layer) → argues from a durable view, not the news cycle.

Conversation tips

→ Come in with a specific technical or structural question about AI infrastructure — he trained as a networking researcher and lights up at stack-level analysis, not use-case storytelling.
→ Reference one of his pieces by argument, not just title — he has written variations of the 'where value accrues in AI' thesis across multiple venues and can tell if you've actually read it.
→ Don't pitch him on AI regulation as a binary (regulate / don't regulate) — his public position is that policy should be evidence-based, so lead with data if you go there.
→ The Nicira story is a genuine inflection point in his arc — asking about the technical insight that made SDN possible (not just the outcome) will resonate more than the $1.26B acquisition number.

Toolbox

Openers

Open on his 2025 'this feels like 1996' framing from The Generalist — it's his sharpest public thesis right now, and asking what specifically maps to 1996 (and what doesn't) gives him something to sharpen.
Reference the Nicira founding: he co-built the company that proved software-defined networking was real, then watched VMware absorb it — that operator history shapes every infrastructure bet he makes now.
His late-2024 Fortune and TechCrunch pieces on AI regulation are a clean entry point: he's publicly arguing that existential-risk framing is distorting policy, which is a concrete, contested position worth probing.

Discovery questions

In your 'where value accrues in AI' thesis, you and Sarah Wang have argued the infrastructure layer captures durable economics — how does that hold up now that model costs are compressing so fast?
When you built Nicira, SDN was considered an academic idea that wouldn't survive contact with enterprise buyers — what's the equivalent contrarian infrastructure bet in AI that most investors are still dismissing?
You've written that AI regulation should be grounded in evidence rather than existential angst — what evidence would actually change your view on where regulation is warranted?

Avoid

Don't lead with AI hype or broad 'AI is changing everything' framing — he is on record arguing that most AI discourse is unmoored from technical and economic evidence, and he'll disengage from vague enthusiasm fast.

Make it yours