a16z × Google Cloud: Inside the AI-Native Venture Stack

Slush investor pass holders only. Venture capital is undergoing its most radical architectural shift in decades. The era of manual CRM entry, surface-level scraping, and fragmented SaaS tools is giving way to proprietary, AI-native platforms engineered for speed, network intelligence, and institutional memory. Join a16z and Google for Startups during Slush for an off-the-record, technical deep dive into how modern venture infrastructure is actually built and deployed at scale. We are joined by Nathaniel Barling (linkedin.com ↗), who leads product and engineering initiatives across a16z’s Applied Software Group. Nathaniel is responsible for architecting the proprietary software and machine learning infrastructure that powers a16z’s deal sourcing, investment decision workflows, and global portfolio operations. Together, we will peel back the curtain on what it takes to build a modern, AI-first stack—and how pushing foundation models into mission-critical workflows reveals the true bottlenecks separating hype from enterprise viability. What We’ll Cover Architecting the AI-Native Venture Stack (Build vs. Buy): How top-tier funds navigate the trade-offs between stitching together off-the-shelf tooling (Affinity, Harmonic, Clay) versus deploying dedicated internal engineering teams to build custom infrastructure. Beyond Surface Metrics to Qualitative Intelligence: Moving past vanity metrics (GitHub stars, LinkedIn headcount shifts) to orchestrating LLMs capable of assessing velocity, technical depth across codebases, and nuanced customer sentiment. Unlocking Years of Institutional Memory: How to effectively query decades of passed deals, partner memos, and board decks to answer the hard questions: “What did we believe about this market three years ago, and which core assumptions broke?” Algorithmic Portfolio Acceleration: The engineering behind matchmaking: connecting founders with the exact Fortune 500 decision-maker, strategic enterprise customer, or VP of Engineering candidate across a network of hundreds of thousands of nodes. Ground Truth from the Frontier: How building internal, production-grade AI systems surfaces the real friction points in context windows, inference latency, agent orchestration, and model reliability—giving investors a clear lens on which AI startups are solving foundational problems. Who Should Attend Venture Capital Partners, Principals, and Platform/Engineering leads looking to modernise their fund’s operational stack. Technical leaders and architects interested in real-world LLM deployment, retrieval, and internal intelligence engines. Please note: this event will take place within the ticketed area of Slush.