Raghu Raghuram is a managing partner at Andreessen Horowitz and a general partner on its Growth and Infrastructure teams. Formerly VMware’s CEO, he writes and speaks about infrastructure strategy, organizational scaling, and the operating work that follows product-market fit.[1]
Start with What VMware’s former CEO says you should prioritize after product-market fit →
Education
Links
a16z
X Managing Partner; General Partner, Growth and Infrastructure · Andreessen Horowitz
Background
Career
Raghuram spent roughly two decades at VMware, moving through product leadership, management of its software-defined data-center business, and the COO role before becoming CEO. His current a16z biography describes leading the company through its Dell spinout and acquisition by Broadcom. Earlier roles included product management at Netscape, AOL, and Bang Networks. He joined a16z in 2025 to help invest in companies and run the firm.[1][2]
Ideas
Operating after product-market fit
Raghuram argues that growth changes the founder’s job. Hiring brings different assumptions about work, so culture needs an explicit framework. Scaling requires choosing priorities and making them usable by people who were not present at the founding. He connects those decisions to his experience moving VMware from a successful product toward a broader platform.[2]
Specialization and production
With Sarah Wang, Raghuram argues that sufficiently large, stable computing workloads reward specialized hardware. Their essay applies that recurring pattern to AI inference and their investment in Etched. It emphasizes the complete system around a chip, including interconnects, memory, power, cooling, and manufacturing. Production capacity becomes part of the product because a promising architecture has little economic impact until customers can run it at scale.[3]
Tools, writing & talks
Writing, research & talks
Start here
A conversation and edited operating advice on culture, priorities, and scaling an infrastructure business.
Read the post-product-market-fit discussion for practical questions about culture and priorities. It is accessible without a technical background and gives context for his later infrastructure investment essays.
Featured operator and speaker; article by Emma Janaskie.[2]Pat Gelsinger discusses AI chip design, why silicon cycles stay long, HBM and packaging limits, optical interconnects, energy as the economic bottleneck, and a VMware-style layer for agent infrastructure.
Catalog workOriginal sourceAWS CEO Matt Garman discusses reshaping cloud for agent-driven infrastructure, reserving scarce GPUs for startups, the Trainium and Graviton bet, AI capex and bubble questions, and what enterprises need to trust autonomous agents.
Catalog workOriginal sourceAn investment thesis connecting specialized inference systems with production and deployment at scale.
Co-author with Sarah Wang.[3]