Velaura AI Launches Titan Core™ for Enhanced Power Efficiency in AI Accelerators
Velaura AI introduces Titan Core™, a silicon design and IP platform that offers up to 2x lower power consumption for AI accelerators. This innovation allows for significant cost savings in data centers and supports increased deployment of AI infrastructure amid rising global electricity demand. Titan Core™ has been validated with over 30 million ASICs in production and integrates seamlessly into existing designs, promising substantial improvements in total cost of ownership.

Velaura AI has launched Titan Core™, a silicon design and IP platform that enables AI accelerators to achieve up to 2x lower overall chip power, resulting in significant savings of up to 500W on typical GPUs or XPUs. This technology addresses the increasing constraints of power availability for hyperscaler data centers, which face projected AI-related electricity demands exceeding 945 TWh/year by 2030.
Titan Core™ reduces energy consumption for essential matrix multiplication operations by 2-4x, translating to approximately $1,300 in electricity savings per XPU over three years. The platform is built on advanced low-voltage design techniques proven in over 30 million ASICs and integrates easily into existing SoC architectures.




Comments