eXistential AI develops AI algorithms optimized for neuromorphic chips, enabling energy-efficient computing solutions. By focusing on explainable AI, they ensure transparency and interpretability of models, aligning with ethical guidelines and AI regulations.
Funding
Funding not disclosed
Founders
Product
Problem
Traditional AI systems often consume significant power, making them unsuitable for portable medical devices that require continuous monitoring and long battery life. Furthermore, the lack of transparency in AI decision-making processes can hinder trust and regulatory compliance in healthcare applications.
Solution
eXistential AI (eXAI) offers energy-efficient AI processing solutions optimized for edge deployment in medical devices. Their BIOT chip utilizes brain-inspired computing principles, such as compute-in-memory and event-based processing, to minimize power consumption while enabling real-time analysis of biosignals. The chip also supports explainable AI (XAI), providing traceable AI processing paths that enhance confidence in device decisions and simplify compliance with AI regulations. By bringing advanced AI processing directly onto the device, eXAI reduces the need for cloud connectivity, further decreasing energy usage and ensuring data privacy.
Target Audience
eXistential AI primarily targets medical device manufacturers developing portable devices that require continuous monitoring, real-time processing, and explainable AI capabilities.
Features
- Ultra-efficient edge processing using compute-in-memory architecture
- Event-based processing that activates spiking neural networks (SNNs) only in response to essential signals
- Explainable AI (XAI) for traceable processing paths and simplified AI Act compliance
- Real-time processing for continuous monitoring and rapid diagnostics of multiple biosignals
- Reduced reliance on cloud connectivity for lower energy consumption and enhanced data privacy
- Compatibility with portable medical devices, including diagnostic equipment and wearable health monitors