Testing SūmerBrain on NVIDIA DGX Station with GB300 versus GH200, mean response time dropped from 4.37 seconds to 2.67 seconds.

SūmerSports, a technology company empowering the evolution of sports with AI, today announced that it has successfully tested and validated its SūmerBrain AI engine on NVIDIA DGX Station™ powered by the NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip, demonstrating significant performance gains for real-time decision-making environments.

SūmerBrain processes complex tracking data, game context, and historical outcomes to deliver actionable insights to teams and leagues. The platform can be used by a coaching staff for in-game decision making, as well as by a front office to improve the fan experience in real time, meaning processing speed is a critical factor in delivering results. SūmerSports evaluated the performance of SūmerBrain on DGX Station with GB300 and identified it as a recommended architecture for local AI development and deployment low-latency use cases, including real-time game analysis, in-game decision support, and a plethora of other live experiences to serve fans and teams, all while guaranteeing the security and compliance that businesses operating in the sports industry require.

Testing SūmerBrain on NVIDIA DGX Station with GB300 versus GH200, mean response time dropped from 4.37 seconds to 2.67 seconds, and across the full range of query complexity, from 9.19 seconds to 4.69. These record-breaking performance benchmarks demonstrate the potential for a more responsive user experience and scalable foundation for SūmerSports to expand its AI solutions across the professional sports market.

"We couldn't be more excited about these results. In the high-stakes world of professional sports, where every decision matters, our AI engine running on the fastest platform in the industry is a difference-maker,” said Lorrissa Horton, CEO of SūmerSports. “Working with NVIDIA, we’ve proven that sports businesses can use AI to quickly make better decisions, surface deeper insights, and compete like never before."

As NVIDIA powers many of the world’s leading OEM systems, this collaboration provides a foundation for SūmerSports to deliver high-performance AI solutions across on-premesis or cloud environments, positioning the company to expand its reach through additional infrastructure and deployment partners. While SūmerSports’ technologies run across multiple NVIDIA architectures, the GB300 is an example of how the latest AI technology running on the latest hardware can enable a new generation of experiences that feel like they’re happening in real-time.

SūmerSports will use the performance enhancement uncovered through its product testing collaboration with NVIDIA to scale its SūmerBrain framework. This AI engine synthesizes real-time and historical data into actionable intelligence, with active proof-of-concept initiatives currently underway across several professional sports organizations and in the financial sector.

In addition to significant improvements in query processing speed, SūmerSports’ collaboration with NVIDIA revealed advantages including:

  • Hybrid Data Flexibility: The GB300 platform offers high-speed integration of both public and private data sets, allowing for comprehensive analysis without compromising performance. Designed for high-stakes environments, the framework enables organizations to synthesize proprietary data in a secure, isolated infrastructure that prevents exposure to the open internet.
  • Information Security: Built with a privacy-first approach, the platform enables organizations to work across public and proprietary data in a secure environment without exposing sensitive information to the open internet.
  • Full-Spectrum Operational Impact: The architecture provides a 360-degree solution for the sports ecosystem, with applications extending from player performance and team operations to enhancing the fan experience. 

“Our work with NVIDIA is foundational to how we scale SūmerBrain across the sports landscape,” said Shawn Butler, SVP of Sales and Strategic Alliances at SūmerSports. “We're excited to deliver deeper insights that help teams compete at a higher level — while giving sports organizations the real-time tools they need to elevate fan engagement and drive meaningful revenue growth.”

“Our technical testing covered a diverse set of real-world workloads, and the results indicate broader performance headroom across language, vision-language, and speech-to-text systems to help reduce friction in AI-driven workflows,” said Hashem Alsaket, Principal AI Engineer, SūmerSports.