Graph-Driven XAI: Rethinking AI Infrastructure Cost with GPU and CPU
A Talk by Ricky Sun (Founder & CEO, Ultipa)
About this Talk
A Case Study in Balancing Performance, Cost, and ESG Using Graph Analytics
In today’s AI landscape, GPUs are often seen as the go-to solution for intensive computing, but at Ultipa, we’ve proven that there’s a better way. With our unique CPU-centric architecture—leveraging both X86 and ARM platforms—we’ve achieved 5-10x performance improvements compared to traditional GPU-based systems, all while lowering total cost of ownership (TCO) by 70% and cutting energy consumption by 75%.
This isn’t just theory; our architecture is already powering real-world applications for global customers, including financial institutions, revenue agencies, supply chain enterprises, and organizations undergoing digital transformation. Looking ahead, we are partnering with CPU, FPGA and ASIC manufacturers to further optimize costs and energy efficiency while pushing the performance envelope even further. This approach ensures superior TCO, ROI, and time-to-value (TTV), making it a future-proof solution for AI infrastructure.
Join us to learn how our graph-augmented XAI framework is reshaping the AI infrastructure landscape, delivering unmatched performance, sustainability, and value to our customers.
About The Speakers
Founder & CEO, Ultipa
A serial entrepreneur, world-class high-performance storage and computing system expert, started his career in the heart of Silicon Valley with his professor a year before his graduation from Santa Clara University. Over the past 20+ years, he went through 3 M&A, while Ultipa is his 4th venture.
We’re a cloud tech company that provides organisations around the world with computing infrastructure and software to help them innovate, unlock efficiencies and become more effective. We also created the world’s first – and only – autonomous database to help organise and secure our customers’ data.
