Principal at Kleiner Perkins
Check size: No personal check range or signing authority is published. Nadia's current portfolio includes early and growth investments; individual company round sizes are not a personal range.
Nadia invests across enterprise applications and infrastructure, from AI compute and networking to vertical AI in healthcare, finance, sales, coding, and productivity. Her official portfolio spans both early and growth stages and emphasizes systems that meet domain-specific standards—clinical evidence, financial auditability, low-latency media, safe autonomy, or enterprise context—rather than generic AI capability alone.
Start with the domain-standard outcome, not a general AI claim. Show proprietary context, workflow integration, evaluation, reliability, latency and cost, auditability or clinical evidence where required, security, and production adoption. Explain what specialized architecture or distribution makes the product hard for a horizontal model provider to absorb.
AI products grounded in proprietary context and domain trust, teams that meet exacting workflow standards, infrastructure that removes a real scaling bottleneck, strong production adoption, and companies whose technical advantage translates into better decisions, safer care, higher productivity, or reliable autonomy.
Her published theses imply caution toward generic assistants without domain context, financial AI without citations or auditability, medical AI that speculates rather than anchors answers in evidence, consumer-like demos that cannot satisfy enterprise reliability, and autonomy that has not progressed from concept to safe commercial deployment.
Define the evidence, audit trail, permissions, reliability, formatting, safety, and escalation standards a vertical AI product must meet before professionals can delegate real work.
Measure where horizontal models fail on proprietary context, latency, citations, precision, workflow completion, or regulation, then test whether the startup owns a durable fix.
Target repetitive work that constrains a scarce expert workforce, automate it safely, and verify that human capacity shifts toward reasoning, decisions, and relationships rather than creating new oversight burden.
Look for repeated commercial use, safety and reliability evidence, broad integration, and behavior normalization—the signals that an emerging technology has crossed from demonstration into infrastructure.
“Every output meets the auditability and citation standards institutional clients demand.”
— https://www.kleinerperkins.com/perspectives/rogo-the-ai-platform-for-global-finance/
“This is no longer about proving a concept.”
— https://www.kleinerperkins.com/perspectives/waymo-the-infrastructure-of-autonomy/
Financial agents need workflow-complete outputs, source auditability, precise formatting, and institutional trust rather than generic text generation.
Autonomy becomes infrastructure when commercial scale and superior safety evidence make driverless service an ordinary, trusted behavior.
Clinical AI must synthesize current medical literature quickly while grounding every answer in authoritative evidence at the point of care.
Generative AI can automate repetitive healthcare work so scarce humans focus on higher-order reasoning, decisions, and patient connection, expanding care supply.