Self-Host Frontier AI: Running Open-Weight Models on AWS

The open-weight model revolution is here frontier models matching or beating closed APIs on benchmarks, and you can run them on your own infrastructure.

About this event

This 90-minute roundtable and workshop focuses on the technical requirements for self-hosting open-weight frontier models on AWS infrastructure¹. Attendees will examine GPU memory calculations, throughput management, and cost comparisons between managed endpoints and self-hosted clusters¹.

The session covers deployment strategies for models including GLM, Kimi K3, Gemma, DeepSeek, and MiniMax M3¹. Participants will review decision frameworks for production-grade serving and methods for implementing day-zero model support¹.

Who is in the room

The event is designed for machine learning and AI engineers responsible for building inference pipelines, as well as platform teams evaluating infrastructure management strategies¹.

How to get in

Interested individuals must RSVP via Partiful¹. The event is scheduled for October 16, 2026, at 1:00 PM¹.

Why it matters

This event is relevant for teams deciding between managed API providers and self-hosted infrastructure to maintain data sovereignty and control¹. It provides a technical guide for scaling models from single-GPU setups to large-scale production environments¹.

Practical

The event takes place in Santa Monica as part of LA Tech Week 2026¹.

Sources

• Self-Host Frontier AI: Running Open-Weight Models on AWS (tech-week.com)