Matt Bornstein

Matt Bornstein is a General Partner on the infrastructure investing team at Andreessen Horowitz, where he focuses on artificial intelligence, machine learning, and data systems. He joined a16z around 2019, and in January 2026 he and colleague Alex Immerman were elevated to General Partner. Over his tenure he became one of the firm's most visible voices on the AI infrastructure stack, co-authoring widely-circulated essays such as "Emerging Architectures for LLM Applications," "Who Owns the Generative AI Platform?," and the "AI Canon" reading list that helped shape a16z's data and AI architecture thesis.

Before a16z, Bornstein was a Principal at the seed-stage venture firm Blumberg Capital (roughly 2013–2019). Earlier, he spent time on the business operations team at LinkedIn (2012) and worked as a management consultant at Monitor Deloitte (2007–2011). He also has hands-on founder experience: he co-founded two bootstrapped startups, including Smarter Feedback, an HR performance-management SaaS platform, and DormSnacks, a dorm-delivery service at Brown University where he served as CTO.

Bornstein holds an MBA from Harvard Business School and a BS in mathematics from Brown University. a16z frequently frames his perspective through the lens of his being a "former founder and CTO," which informs his hands-on, technically deep approach to backing infrastructure companies.

• Invests primarily in AI/ML, data infrastructure, and developer tools — the picks-and-shovels layer of the modern software and AI stack.
• Operates within a16z's institutional structure, so check sizes range from early seed rounds into growth, but his sweet spot is early-stage (seed/Series A) infrastructure.
• Looks for deeply technical teams building foundational systems — data platforms, model-serving/inference, evaluation and labeling, and the tooling that productionizes AI applications.
• His public writing signals the thesis: he is interested in who captures value in the generative AI platform (infrastructure and models vs. apps), emerging architectures for LLM applications, data labeling and expert evals as an "unsung hero" of AI, and the need for continual learning in models.
• Founder-friendly and technically engaged — a16z highlights his background as a former founder and CTO, suggesting he values builders who understand the systems layer firsthand.