elastix.ai

ElastixAI is a Seattle-based startup developing a next-generation AI inference platform to optimize the performance and efficiency of large language models (LLMs). The company, currently in stealth mode, was founded in early 2025 by veteran engineers from Apple, Xnor, Meta, and Waymo. The founding team includes CEO Mohammad Rastegari, CTO Saman Naderiparizi, and Mahyar Najibi. Rastegari and Naderiparizi were key figures at Xnor, an edge-AI startup acquired by Apple for approximately $200 million in 2020.

The company's software-oriented platform is designed to reduce the cost and complexity of running AI models. It addresses the critical AI inference stage, which is how trained models generate responses. The technology is hardware-agnostic, aiming to improve LLM efficiency on a wide range of hardware configurations, from edge devices to large-scale cloud deployments. This allows customers, from hyperscalers to enterprises, to configure and fine-tune their infrastructure for specific use cases, delivering AI inference at scale and a lower cost.

In May 2025, ElastixAI announced it had raised $16 million in a seed funding round. The round was led by Bellevue-based venture capital firm FUSE, with participation from other investors including Catapult, Tyche, Liquid 2 Ventures, and DNX Ventures. The business model targets a significant market by providing a flexible solution for companies integrating AI into their daily operations. By focusing on software-driven optimization, ElastixAI aims to become a key infrastructure provider in the growing AI inference market.

Keywords: AI inference platform, large language models, cost-efficient AI, model optimization, edge AI, cloud deployment, hardware-agnostic AI, scalable inference, GenAI infrastructure, LLM efficiency, AI software, enterprise AI, hyperscaler solutions, Mohammad Rastegari, Saman Naderiparizi, Mahyar Najibi, FUSE, Xnor