Ex-NEA partner launches Premise VC, betting on early-stage founders over big fund model

Vanessa Larco spent nearly eight years as a partner at New Enterprise Associates (NEA), where she sat on the firm's investment committee and led deals in enterprise software, developer tools, and consumer technology including Evident, Kindred, Cleo, Greenlight, and Mejuri. She also served as a board observer at Robinhood ahead of its 2021 IPO. Before venture capital, Larco built a career as a product leader and founder: she earned a computer science degree with honors from the Georgia Institute of Technology, worked at Microsoft on Xbox and Kinect V1, led product teams at Twilio and Box, and founded and sold an app development startup.

In early 2025, Larco co-founded Premise VC with Mercedes Bent to target a market gap she identified at NEA. At a firm deploying between $3 billion and $6 billion annually, writing $2 million checks is never a portfolio priority relative to larger deployments. Larco noticed that founder preferences had shifted: one founder told her, "I want my investor at every round to feel like the check size hurt—that it's a big percentage of their fund—because that's how I know I'm going to be a top priority when push comes to shove." The Silicon Valley Bank collapse accelerated this realization. When SVB failed, founders called their entire cap tables asking for emergency payroll funding. Every VC received dozens to hundreds of calls, forcing firms with large portfolios to triage. Founders who weren't helped saw exactly where they ranked on their investors' priority lists. Word spread rapidly through founder networks, and emerging specialized funds began winning competitive deals against mega-funds—reversing a decades-old pattern. Premise VC writes checks of $500,000 to $3 million at pre-seed and seed stages exclusively.

Larco evaluates founders not primarily on their initial idea but on founder potential and core attributes. She and Mercedes Bent identified seven founder attributes across their best investments at previous firms and look for founders who are world-class in at least two of them. During diligence, they speak with founders one to three times a day for three to five days, along with extensive reference and back-channel checks. Most investments are in cities where they have strong networks: San Francisco, New York, and Atlanta. Larco emphasizes that at pre-seed and seed stages, the initial idea is unlikely to match the company's shape in five to ten years; the evaluation focuses on whether the founder can find the right market and product fit. One surprising attribute on her list is what she calls "urgently dissatisfied"—founders who can seem disagreeable, prioritizing goals over people-pleasing, with standards that push teams beyond what they think possible. These founders are not ego-driven but relentlessly focused, with high bars for themselves and their teams.

On AI specifically, Larco looks for teams that make products dramatically faster, cheaper, or easier than existing options—not merely API wrappers. She is willing to back wrapper companies that deliver a 50% cost reduction or reduce a five-hour weekly task to five minutes, but founders must understand the underlying mechanics. Technical founders optimize models for specific features; less technical founders often use a single model for everything, creating risk if that model becomes expensive or degrades. She notes that when founders are wedded to a single model and don't know how to innovate beyond it, the moat is fragile. However, she acknowledges that early AWS criticism—that anyone could build a company on AWS without owning servers—proved unfounded; many great companies were built on cloud infrastructure. The same pattern held for mobile and may hold for AI-first startups built on third-party models.