Rebellions Advances Peta-Scale Inference with proteanTecs Deep Data Monitoring

HAIFA, ISRAEL – October 27, 2025 – proteanTecs®, a global leader in deep data solutions for electronics health and performance monitoring, announced today that Rebellions, a cutting-edge AI semiconductor company, has adopted proteanTecs’ embedded lifecycle monitoring analytics for its purpose-built AI accelerators, designed for inference at scale.

‍Rebellions’ AI accelerators are engineered for the new era of large-scale AI. Supporting workloads from LLMs to multi-modal applications, the company’s advanced chiplet architecture, ultra-high-bandwidth HBM3E integration, and optimized software stack enable hyperscale deployment with breakthrough energy efficiency and performance. Rebellions is a leading semiconductor company delivering both chiplet-based designs and HBM3E into production, powering peta-scale mixture-of-experts (MoE) inference and setting new benchmarks for AI infrastructure.

‍By embedding proteanTecs’ monitoring IP and leveraging its ML-driven analytics, Rebellions gains real-time chip telemetry insights into their systems’ performance and health. This enables significant power reduction, faster time-to-market, and enhanced quality and reliability.  

‍“Our mission is to deliver next-generation AI processors that balance performance and cost-efficiency,” said Jinwook Oh, Co-founder and CTO at Rebellions. “proteanTecs gives us deeper visibility into our chips, from product ramp through volume production and into datacenter deployment. Their deep data monitoring solutions allow us to push performance-per-watt even further while ensuring resilience at scale.”

‍“As AI adoption accelerates globally, balancing performance gains with energy efficiency and system robustness is critical for cloud-scale deployments,” said Sanjay Lall, Chief Revenue Officer at proteanTecs. “Rebellions is driving a new standard in AI acceleration with its high-performance, high-bandwidth, chiplet-based designs. We’re proud to support their innovation as they successfully scale to meet the most demanding workloads.”