Martin Casado

Who they are

Martin Casado is a General Partner at a16z — co-founded Nicira Networks, the networking company he built as CTO before its acquisition, then earned a Stanford PhD in networking before moving into venture.

Person

Martin Casado did his Master's at Stanford in 2005 and stayed through a PhD, finishing in 2008 — the dissertation work seeded what became Nicira Networks, which he co-founded in 2007 and ran as CTO until its acquisition in 2012. At VMware post-acquisition he rose to GM Networking and Security and SVP and Fellow — an unusually technical title for a business-unit leader, signaling that his operator credibility stayed intact through the integration. He joined a16z in April 2016 to lead the infrastructure practice. The through-line is a founder-turned-operator-turned-investor who has lived on the infrastructure layer at every stage: built it, scaled it through a large acquirer, and now bets on who builds it next. He is a prolific public voice — a16z podcast episodes in 2024 on AI value accrual, SaaS pricing disruption, and the future of VC; a 2025 appearance on The Generalist where he compared the current AI moment to 1996; and published takes in Fortune and TechCrunch pushing back on evidence-free AI regulation. His content is consistently adversarial toward regulatory overcaution and analytically focused on where economic value concentrates in AI stacks.

Company

a16z closed a $15 billion multi-fund raise in January 2026 — its largest haul to date — with a stated thesis of backing AI companies that replace operational drag. In May 2026, the crypto arm raised a separate $2.2 billion fifth fund dedicated entirely to crypto entrepreneurs, focusing on practical applications like stablecoins and financial infrastructure rather than speculation, and promoted its crypto CTO to General Partner. The firm manages $90 billion in assets under management as of early 2026, with a portfolio that has grown to 1,159 companies including 182 investments made in 2025 alone. Recent deal activity includes leading a $35 million Series A for AI startup Lassie, seed funding for domestic manufacturing startup Westmag, and a funding round for blockchain infrastructure provider Digital Asset Holdings targeting a $2 billion valuation. The firm has also rebranded its Investor Relations function to Global Partnerships, signaling a push toward sovereign and institutional LPs at international scale.

Market

a16z competes directly with Sequoia, Accel, General Catalyst, Founders Fund, Bessemer, NEA, Index, and GV for the most competitive deals across AI, crypto, defense, and frontier tech. Its 'American Dynamism' thesis — investing over $1 billion in aerospace, defense, and public safety aligned with national security interests — differentiates it from peers and reflects geopolitical tailwinds around domestic manufacturing and Pentagon priorities. Regulatory uncertainty in crypto remains a live risk; a16z is actively lobbying for the US CLARITY Act to reduce policy risk premium across its crypto portfolio.

Network

Martin's most visible internal collaborator is Marc Andreessen, with whom he co-presents on AI themes including a 2024 episode on AI saving the world. Sarah Wang is his co-presenter on AI value accrual analysis. Alex Rampell, co-founder of Affirm and a16z General Partner, is active on the fintech and AI side — he joined the board of Lassie in a recent a16z-led round.

Marc Andreessen· Co-founder and General Partner, a16z
Sarah Wang· General Partner, a16z
Alex Rampell· General Partner, a16z

How they likely show up

PhD from Stanford in networking followed by founding and CTO-ing Nicira → he leads with technical depth first; credibility for him comes from first-principles understanding, not pattern matching.
Rose to SVP and Fellow at VMware after acquisition → comfortable operating inside large institutional structures without losing independent technical identity.
Long tenure at a16z since April 2016 → thinks in multi-year investment cycles; not a flip-the-portfolio operator.
High volume of public appearances in 2024–2025 across podcasts, TechCrunch, Fortune, and Politico → comfortable being visible and taking positions that invite pushback; likely engages best when the conversation has a real argument in it.
Content themes cluster tightly around AI infrastructure, value accrual, and regulatory skepticism → he has developed and published specific views; he will want to debate, not just discuss.
Compared the current AI moment to 1996 in a public interview → frames his investment thesis in historical analogies; responds well to structured comparisons.

Conversation tips

→ Come in with a specific view on where value accrues in the AI stack — he has published his own thesis on this and will immediately test whether you've thought about it independently.
→ Reference his Nicira work or the VMware integration if you want to ground the conversation in operator experience; it's the credibility anchor he carries into every infrastructure bet.
→ Don't expect a soft regulatory conversation — he has been publicly and repeatedly critical of evidence-free AI regulation; if you have a policy angle, bring evidence.
→ Engage with the 1996 analogy he used on The Generalist; asking where he thinks the analogy breaks down is a sharp way in.
→ Be specific about the infrastructure layer you're discussing — he distinguishes between where AI value accrues at the model, infrastructure, and application layer, and conflating them will signal you haven't done the work.

Toolbox

Openers

Open on Nicira — he co-founded it in 2007 straight out of his Stanford PhD work, ran it as CTO through the acquisition, and it's the hands-on infrastructure credential that every subsequent bet traces back to.
Reference his 2025 Generalist interview where he said 'this feels like 1996' — a specific and contestable claim that he put on record, and a natural way to get him talking about where the analogy holds and where it doesn't.
Bring up the a16z podcast episode on AI upending SaaS pricing — he has a specific, published view on how AI disrupts the per-seat model, and it's directly relevant to any infrastructure or application company in the room.

Discovery questions

You've argued that AI value will concentrate at specific layers of the stack — how has that thesis shifted, if at all, since you first published it in 2024?
Running Nicira through the VMware integration gave you a rare view of what happens to infrastructure companies post-acquisition at scale — how does that experience shape which infrastructure bets you make today?
You've been vocal that most AI regulation is built on existential angst rather than evidence — what would evidence-based AI policy actually look like to you, and which proposals come closest?

Avoid

Don't come in with vague AI optimism or hype framing — he explicitly pushes back on non-evidence-based AI narratives and will lose interest fast if the conversation isn't analytically grounded.

Make it yours