AI Reliability: A GPU - Native Framework for Real-Time Spectral Telemetry and Anomaly Sequencing in Data Centers and Beyond

Authors

Amos Njeru, Robotics Reliability Research Lab, Kenya

Casam Njagi, Chuka University, Kenya

Ruth Wario, University of the Free State, South Africa

Rosa Njagi, Grand Canyon University, United States

Published in

IST-Africa 2026 Conference Proceedings

ISSN: 2576-8581

ISBN: 978-1-905824-76-2

DOI: https://doi.org/tbc

Publisher

IST-Africa Institute and IIMC International Information Management Corporation Ltd

Published in Ireland

Abstract

Modern AI data centers operate as living digital ecosystems whose pulse is measured by continuous GPU telemetry across compute, cooling, and power domains. Yet, current monitoring systems remain threshold-based, detecting anomalies only after degradation occurs. This study introduces Eigenscan linear algebra framework, a GPU-native diagnostic framework developed in CUDA's computational language using cuDF, cuBLAS, and cuSolver. The framework executes eigen decomposition up to 50× faster than CPU systems, sequencing compute stability as a dynamic spectral health map. Synthetic data mirroring NVIDIA's Data Center GPU Manager (DCGM) and NVML APIs were generated to simulate operational telemetry. Each GPU signal temperature, voltage, and memory utilization represents a digital heartbeat. These were analyzed to detect decay corridors where micro-drifts evolve into systemic instability. The system timestamps each event into an Anomaly Decay Clock, enabling predictive orchestration and early intervention. Results demonstrate that the first eigenmode explained over 85% of system trust variance while GPU acceleration achieved millisecond-level performance. This approach advance's reliability, compliance, and self-awareness in Physical AI, positioning Spectral Telemetry as the missing stability layer in next-generation compute ecosystems.

Keywords

Spectral Telemetry; Anomaly Sequencing; Eigen Decomposition; Covariance Drift; Trust Health Index; Mahalanobis trust corridor; Cyber-Physical Systems; Predictive Maintenance; GPU Acceleration

Cite this paper

A. Njeru, C. Njagi, R. Wario and R. Njagi (2026) "AI Reliability: A GPU - Native Framework for Real-Time Spectral Telemetry and Anomaly Sequencing in Data Centers and Beyond", IST-Africa 2026 Conference Proceedings, Miriam Cunningham and Paul Cunningham (Eds), IST-Africa Institute and IIMC, 2026, ISSN: 2576-8581, ISBN: 978-1-905824-76-2, https://doi.org/tbc

Access Paper

Click here to access this paper.

IST-Africa Community Members can download papers by logging into their existing account.

If you wish to become a member of IST-Africa Community, pleaseregister your profile and you will then receive login details which you can use to access the papers.

Access to papers in the IST-Africa 2026 Conference Proceedings is for INDIVIDUAL USE only. Neither the Conference Proceedings (or individual double blind peer reviewed papers) are to be included in any public or institutional repository or processed in the context of any AI tool or large language model without prior written permission from the Publishers.

While the IST-Africa Institute and IIMC retain copyright of the Conference Proceedings, each author retains copyright to their individual paper. No paper or part thereof may be reproduced without the written permission of both the Conference Proceedings Publishers and appropriate author(s).