Zabie Elmgren

There is a version of Zabie Elmgren's story that starts at Harvard. It starts with competitive rowing at the Head of the Charles Regatta - sitting bow, three, seven, switching seats in the shell, learning to move as one unit. But the more interesting version starts a little later, around 2016, when she joined the founding team of a company called Human Capital while still an undergraduate. Most people at that point were writing essays. Zabie was helping to build a venture fund.

Human Capital began as a recruiting firm. Zabie helped turn it into a $600 million asset manager. Along the way, she personally placed some of the first fifty engineers at Brex, Robinhood, Snowflake, and Nuro. Snowflake alone went on to pull off the largest software IPO in history at that time. This is not context she often leads with. It is the kind of track record that speaks for itself if you know where to look.

What makes her unusual among venture capitalists is the specific lens she carries from that recruiting background. She has sat across from hundreds of founding teams not as a check-writer evaluating a pitch, but as the person responsible for finding the humans who would actually build the thing. She knows which founding teams hold together under pressure before they know it themselves. "Recruiting is not something to delegate," she has said. "It is the job." That conviction runs through everything she does.

After Human Capital, Zabie moved across the table. She joined Index Ventures as a Partner in 2021, focused on early-stage B2B and B2C software. Index, with offices on both sides of the Atlantic, was a different operating environment - one of the most respected early-stage firms in the world, with a portfolio that had produced Slack, Dropbox, and Roblox. She held her own. Her investments included Anduril, the defense-technology company now valued in the tens of billions, Applied Intuition, the autonomy software platform used by automotive and defense clients, and Verkada, the physical security company that has become infrastructure for thousands of buildings.

There is a thread running through those three companies: they all touch the physical world. They do not live entirely in the cloud. Anduril builds autonomous systems for national defense. Applied Intuition helps vehicles sense and navigate reality. Verkada puts sensors in hallways. This was not a coincidence. It was the beginning of a thesis.

In the fall of 2024, Zabie joined Andreessen Horowitz as a Partner on the American Dynamism investing team. The phrase "American Dynamism" is a16z's shorthand for a thesis about national importance: companies that strengthen critical infrastructure, public safety, defense, manufacturing, and the industrial economy of the United States. It is not the sexiest corner of venture capital. It is, arguably, the most consequential.

Her particular focus within that framework is what she calls physical AI - not AI that generates text or images or code, but AI that is deployed into the physical world. Factories. Supply chains. Industrial systems. Infrastructure you can stub your toe on. The argument she and the a16z team have been making publicly is that the next wave of AI deployment will not happen on screens. It will happen in buildings, on shop floors, in the systems that keep cities and supply chains running. In December 2025, she appeared on a16z's "Big Ideas 2026" podcast laying out the case for physical observability - the idea that the same monitoring and instrumentation that transformed digital software will need to be built from scratch for the physical world.

This is a thesis with real money behind it. In May 2026, Zabie led a16z's $60 million Series A investment in Tessera Labs, a company building AI-native automation for enterprise resource planning - the software that runs procurement, manufacturing, and operations in major corporations. ERP systems are notoriously ancient. Tessera is betting they can be rebuilt for the AI era. Zabie is betting on Tessera.

Zabie studied computer science at Harvard with a secondary in economics, graduating in 2018. She was not a tourist in the engineering curriculum. She rowed on the Harvard-Radcliffe heavyweight women's crew team, competing in multiple seats at major regattas including the Head of the Charles - one of the largest rowing events in the world. She also competed on the Freeride Tour in skiing, which is a separate, arguably more alarming pursuit involving steep mountain faces and no script.

The athletic background is not decorative context. She has said the best piece of advice she ever received as a competitor - "get comfortable being uncomfortable" - applies equally to anyone with genuine ambition. For an investor who bets on defense technology, industrial AI, and companies trying to rebuild American manufacturing, a tolerance for uncomfortable situations is not optional.

Her professional name has always been Zabie, not Elizabeth. Her work email at a16z is zabie@a16z.com. She appears in every professional context, every byline, every conference panel as Zabie. It is a small thing that says something about how she operates: she is not performing a version of herself she thinks the industry expects.

Most venture capitalists who have not been operators themselves tend to underweight the execution difficulty of building a company. Zabie's career started in the mechanics of company-building - not pitching, not writing memos, but actually finding the engineers who would build the first product. She knows what a founding team looks like when it is misconfigured at the skill level. She has co-authored a guide on how to hire a strong founding team, published on a16z, that distills the frameworks she developed placing people into companies like Snowflake before Snowflake was a known quantity.

That perspective shapes her work with portfolio companies beyond just the check. She is involved with how founders build their teams, how they think about culture, how they attract the kind of engineering talent that can execute on hard physical-world problems. This is not generic "value-add" language. It is a specific capability she brings from a decade of actually doing that work.

The companies she tends to back share a quality: they are solving problems that cannot be solved by software alone. Diode automates PCB workflow with AI and real-world production constraints. Tessera Labs is rebuilding ERP systems - the unsexy backbone of industrial enterprise. Base Power Company is building energy infrastructure. These are not companies that get acquired after two years for a talent hire. They are bets on decade-long infrastructure transformations, which is a different kind of patience and a different kind of conviction.

Zabie writes and speaks on two related subjects: how to build founding teams, and why the physical world is the next frontier for AI investment. Her piece co-authored for a16z on how to hire a strong founding team is among the more practical guides available for early-stage startups - it is not abstract advice about culture. It is frameworks derived from having done the work herself, at scale, across dozens of companies.

Her piece "Every Building You've Ever Been In Was Designed By Software Built in 1997" is a more provocative argument - that entire categories of critical infrastructure software are waiting for a Figma-style redesign moment, and that the opportunity is enormous precisely because the incumbents are ancient. This is not a marketing piece. It is the kind of writing that gets forwarded by engineers who have spent years fighting with legacy CAD software.