Mirai Raises $10M to Build On-Device AI Layer
Mirai has raised a $10 million seed round to build an on-device AI capability layer for local inference. The company aims to abstract away the complexity of deploying and maintaining AI models on edge devices. The funding signals investor interest in tools that democratize access to LLM-powered features for developers and end-users, particularly in privacy-sensitive applications.
- The founding team consists of Dima Shevts and Alexey Moiseenkov, the co-founders of the viral consumer AI applications Reface and Prisma, respectively, bringing experience in shipping AI-powered products to millions of users. - The $10 million seed round was led by Uncork Capital, with participation from notable angel investors such as the co-founder of Snowflake, the co-founder of ElevenLabs, a board member of Coinbase, and a former technical leader from Netflix. - Mirai's proprietary inference engine, written in Rust and initially optimized for Apple Silicon, claims to increase AI model generation speed by up to 37% without modifying the model's weights, thereby preserving output quality. - For developers, the company is creating a Software Development Kit (SDK) designed for simple integration, comparable to Stripe's API, enabling features like text summarization and voice processing to be added with minimal code. - The platform architecture includes a workload orchestration layer that can intelligently route requests between the local device and the cloud, allowing developers to leverage on-device processing for speed and privacy while offloading more complex tasks. - Mirai's strategy aligns with a broader industry shift towards on-device AI, as major hardware companies like Apple, Google, and Qualcomm increasingly integrate dedicated AI silicon into their devices to improve performance and reduce reliance on cloud infrastructure. - The company plans to use the funding to expand support beyond Apple Silicon and to cover additional AI workloads, including vision and other multimodal models, aiming to establish on-device processing as the default for many AI applications. - The on-device AI market is projected to reach an estimated $124 billion by 2032, driven by benefits such as significantly lower energy consumption, reduced carbon footprints, and improved data privacy compared to cloud-based processing.
