NVIDIA GTC 2025: WEKA Expands NVIDIA Integrations and Certifications

At NVIDIA GTC 2025, WekaIO, Inc. announced it is integrating with the NVIDIA AI Data Platform reference design and has achieved NVIDIA storage certifications to provide optimized AI infrastructure for the future of agentic AI and reasoning models.

Additionally, the company announced new certifications for the NVIDIA Cloud Partner (NCP) Reference Architecture with NVIDIA GB200 NVL72 and the NVIDIA-Certified Systems Storage designation for enterprise AI factory deployments with NVIDIA Enterprise Reference Architectures.

The company also unveiled its new Augmented Memory Grid capability, which integrates WEKA Data Platform software with NVIDIA accelerated computing, networking, and enterprise software to accelerate AI inference, maximize the number of tokens processed per second, and dramatically increase token efficiency.

Powering AI Agents with WEKA’s High-Performance Storage
NVIDIA AI Data Platform is redefining enterprise infrastructure for the era of agentic AI. It provides a customizable reference design integrating the NVIDIA Blackwell platform, NVIDIA BlueField DPUs, NVIDIA Spectrum-X networking, and NVIDIA AI Enterprise software with enterprise storage to transform data into actionable intelligence. Organizations can now leverage the benefits of the NVIDIA AI Data Platform with WEKA Data Platform software to create a massively scalable, high-performance foundation for enterprise AI that connects AI query agents to business knowledge, and achieves peak AI inference performance and higher accuracy for complex reasoning.

Breaking AI Memory Barrier with WEKA Augmented Memory Grid

AI agents continue to expand autonomous decision-making, complex problem-solving, and adaptive learning capabilities, increasing the need for AI infrastructure that can support longer context windows, expanding model parameters, and growing system memory requirements. With WEKA Augmented Memory Grid, AI models can extend memory for large model inferencing with additional PBs of capacity by 3 orders of magnitude greater than today’s fixed increments of single terabytes. At the same time, the WEKA Augmented Memory Grid can deliver near-memory speed performance at microsecond latencies for faster token processing, enabling unprecedented reasoning outcomes. Key benefits include:

• Faster Time to First Token: When processing 105,000 tokens, WEKA’s Augmented Memory Grid reduced time to 1st token by 41x compared to recalculating the prefill context.

• Optimized Token Processing: Inferencing clusters can achieve higher token throughput across the cluster, lowering the cost of token throughput by up to 24% for the entire inference system.

Advancing Enterprise AI Innovation with New NVIDIA Storage Certifications

WEKApod Nitro Data Platform Appliances have been certified as one of the first high-performance storage solutions for NVIDIA Cloud Partner (NCP) deployments with NVIDIA HGX H200, B200, and GB200 NVL72, to supercharge NCP providers’ infrastructure services for AI developers and innovators. WEKApod appliances deliver high-performance density and power efficiency — a single 8U entry-level configuration can support up to 1,152 GPUs. 

WEKApod Nitro appliances have also achieved the NVIDIA-Certified Systems Storage designation for enterprises deploying AI factories based on NVIDIA Enterprise Reference Architecture guidelines with NVIDIA-Certified Systems. This certification validates that the WEKA Data Platform is compatible with NVIDIA best practices to ensure optimal storage performance, efficiency, and scalability for a wide range of enterprise AI and HPC workloads.

“In collaboration with NVIDIA, WEKA is delivering high-performance AI storage solutions to organizations with the NVIDIA AI Data Platform, tackling data challenges that constrain AI innovation and force compromises in model capabilities and infrastructure efficiency,” said Nilesh Patel, CPO, WEKA. “Just as breaking the sound barrier unlocked new frontiers in aerospace innovation, WEKA Augmented Memory Grid is shattering the AI memory barrier, expanding GPU memory and optimizing token efficiency across the NVIDIA AI Data Platform. This breakthrough will transform AI token economics, enabling faster innovation at lower costs without compromising performance.”

“Enterprises looking to harness the power of agentic AI and reasoning models need unprecedented efficiency and scalability for these demanding workloads,” said Rob Davis, VP, storage networking technology, NVIDIA Corp. “Pairing NVIDIA and WEKA technologies enables AI agents to access and process data with state-of-the-art speed and accuracy during inference.”

At NVIDIA GTC 2025: Attendees can visit the WEKA booth in the GTC Expo Hall to demo the WEKA Augmented Memory Grid capability.

Availability:

• WEKA’s NCP reference architecture for NVIDIA Blackwell systems will be available later this month.

• WEKA Augmented Memory Grid capability will be generally available for WEKA Data Platform customers in Spring 2025.

Resources:
Blog: New Augmented Memory Grid Revolutionizes the Economics of AI Inference Infrastructure   
Blog: WEKA Unleashing AI Reasoning with NVIDIA Blackwell