Webinar
AI Inference: The Compute Layer Powering Everything
Join Meera Oak, Partner at Alumni Ventures, and Senior Associate Lucy Friedmann for a look at why AI inference — not training — is on track to account for 55% of AI-optimized cloud infrastructure spending in 2026 and more than 65% by 2029, and where the venture opportunity is forming across a compute layer racing to serve a plural, multi-model world at a cost that keeps falling.
In this session, Meera and Lucy will unpack why the “which lab wins” question in AI turned out to matter less than expected — because almost no enterprise runs on a single model anymore. They’ll walk through the forces driving that plurality (task-specific model selection, cost discipline as token spend becomes a real budget line, the sovereignty concerns triggered by episodes like the June 2026 export-control restriction on Anthropic’s Fable 5 and Mythos models, and the growing supply of open-weight models from labs like DeepSeek, Qwen, Kimi, and Mistral) and how that demand is reshaping the inference infrastructure stack — from Groq and Etched’s AI-specific silicon targeting the “memory wall,” to serverless inference clouds like Together AI, Lambda Labs, Modal, and Fal, several of which are AV portfolio companies.
Whether you’re an existing AI First Syndicate member sharpening your view of AI infrastructure, or an accredited investor evaluating the sector for the first time, this session will give you a grounded, investor-focused thesis on the layer of AI where value is quietly accruing today. Register to hear directly from Meera and Lucy — and bring your questions to the live Q&A.
Want to access our AI First Syndicate startup deals to see, evaluate, and optionally invest? No cost, no commitment.
Why Attend?
- HomeUnderstand why "which model wins" turned out to be the wrong question — and how enterprises running multiple models in production have made inference infrastructure the real venture opportunity
- HomeSee how cost (~100x decline in LLM inference cost over two years), sovereignty (export controls, data residency), and open-weight models are reshaping who wins in the inference stack
- HomeHear how Alumni Ventures is investing across the inference layer — from Groq's AI-specific silicon and next-gen entrants like Etched, MatX, and Fractile tackling the "memory wall," to serverless inference clouds like Together AI, Lambda Labs, Modal, and Fal
Rated a Top 20 VC Firm by both TIME and CB Insights.
Alumni Ventures is America’s largest venture capital firm for individual investors.
About your presenters
Partner
Meera Oak invests in AI-native infrastructure redefining how software is built. She views this shift as the next cloud moment, a fundamental transition where value moves from manual execution to intelligent orchestration and simplicity. As engineering moves from direct control to higher-level automation, Meera focuses on the foundational layers of compute, data, and agents.
At the investment level, her lens prioritizes agent development infrastructure and AI-native frameworks that favor simplicity over complexity. She seeks out the operating systems for teams made up of both humans and AI agents, with a focus on orchestration, memory, and context coordination. These platforms form the backbone of the next enterprise stack, much like AWS and Salesforce did for the cloud era, by making powerful capabilities feel simple and developer-first.
Meera has built an extensive portfolio of leaders defining the modern AI infrastructure stack across agents, compute, and data. She backed OpenHands, an open source platform for cloud coding agents, and Standard Kernel, which is building a self-improving system that automatically generates and optimizes high-performance GPU kernels as core infrastructure for agentic and frontier AI workloads. Her investments also include Fundamental Research Labs for frontier model research, Atero, a GPU optimization company acquired by Crusoe, and Vamana, where she led AV’s investment in AI-native video infrastructure that makes video computable, searchable, and model-ready.
These companies reflect Meera’s broader strategy to invest in enduring platforms that re-architect critical systems. Her background leading product for cloud-based ERPs gives her the fluency to connect with developers navigating today’s platform shift. Whether backing agentic platforms or security layers, she believes the next generation of venture-scale outcomes will come from rebuilding the infrastructure that underpins the digital world.
Principal
Before joining Alumni Ventures, Lucy built her career launching new products at Amazon, first for AWS and then for Amazon Devices. She graduated with honors from Yale University in 2019 and earned a dual MBA/MA from the University of Pennsylvania’s Wharton School and the Lauder Institute in 2025, where she focused on finance and European venture capital. A lifelong fencer, she competed all four years on Yale’s varsity team. Now happily retired from the sport, she spends her free time hitting tennis balls with friends and supporting the arts, including her volunteer work with American Ballet Theatre’s Junior Council.