Vietnamese
At GTC 2026, IBM and NVIDIA announced a major expansion of their decade-long strategic partnership, focused on advancing AI from pilot projects to full-scale enterprise production. The collaboration addresses several key barriers to enterprise AI adoption, including GPU-native data analytics, intelligent document processing, and secure infrastructure for regulated industries. Together, the two companies aim to deliver a unified stack covering data foundations, optimized hardware, and professional consulting services to help organizations manage fragmented data and meet strict compliance requirements.
IBM Chairman and CEO Arvind Krishna noted that the next stage of enterprise AI relies on deep integration of data, infrastructure, and orchestration. He stated that the partnership is built to provide the tools businesses need to move from AI experimentation to reliable, operational deployment. NVIDIA founder and CEO Jensen Huang highlighted that by embedding CUDA acceleration directly into the data layer, the companies are transforming traditional data processing into real-time intelligence engines.
Accelerating Structured Data Analytics via cuDF and Presto
A core technical focus of the announcement is the integration of NVIDIA’s cuDF library with the Presto SQL engine within IBM watsonx.data. This open-source combination enables GPU-accelerated querying on massive datasets, drastically cutting the time and cost of deriving insights from structured data.
The firms validated this solution in a live production environment with Nestlé’s Order-to-Cash data mart, which processes terabytes of data across 44 global tables. On conventional CPU-based infrastructure, one data refresh cycle took 15 minutes and supported only a few runs per day. After switching to the GPU-accelerated watsonx.data engine, Nestlé reduced query runtime to just three minutes. This delivers a 30X improvement in price-performance and an 83% reduction in operational costs. Chris Wright, Nestlé’s Chief Information and Digital Officer, noted that this capability enables faster decision-making in manufacturing and logistics through near-real-time operational data.
Unlocking Unstructured Data with Docling and Nemotron
To tackle data locked in unstructured formats such as SharePoint sites and content management systems, IBM and NVIDIA introduced a joint solution using IBM Docling and NVIDIA Nemotron open models. Docling standardizes and converts complex documents into AI-ready formats while keeping full traceability to their original sources.
When combined with NVIDIA Nemotron models, the platform speeds up ingestion of multi-modal content. Early tests show significantly higher throughput than many existing open-source alternatives. This approach helps enterprises build a trusted data foundation for AI by making internal knowledge bases more accessible and consistent for automated reasoning.
GPU-Optimized Infrastructure and Sovereign AI
The partnership also extends deeply into storage and infrastructure. NVIDIA has chosen the IBM Storage Scale System 6000 to supply 10 PB of high-performance storage for its GPU-native advanced analytics engines. Certified and validated for NVIDIA DGX platforms, Storage Scale 6000 merges IBM’s high parallel throughput with NVIDIA’s GPU data pipelines to eliminate I/O bottlenecks.
For organizations with stringent data residency and regulatory rules, the pair is exploring integration between IBM Sovereign Core and NVIDIA infrastructure. This effort aims to allow GPU-heavy AI workloads to run entirely within defined geographic boundaries, ensuring governance and compliance without compromising compute performance.
Expanding the AI Stack with Blackwell and Red Hat
IBM announced plans to deploy NVIDIA Blackwell Ultra GPUs on IBM Cloud starting in early Q2 2026, targeting large-scale model training and high-throughput inference. The Blackwell architecture will also be integrated into the Red Hat AI Factory alongside NVIDIA and VPC servers, delivering enterprise-grade controls for data residency.
In addition, IBM Consulting will offer the Red Hat AI Factory with NVIDIA via the IBM Consulting Advantage platform. This is designed to simplify and accelerate data preparation, model development, and deployment workflows. By combining these technologies, IBM and NVIDIA aim to provide a smoother path for enterprises to scale AI across diverse technology environments while maintaining governance and consistent performance.
Beijing Qianxing Jietong Technology Co., Ltd.
Sandy Yang/Global Strategy Director
WhatsApp / WeChat: +86 13426366826
Email: yangyd@qianxingdata.com
Website: www.qianxingdata.com/www.storagesserver.com
Business Focus:
ICT Product Distribution/System Integration & Services/Infrastructure Solutions
With 20+ years of IT distribution experience, we partner with leading global brands to deliver reliable products and professional services.
“Using Technology to Build an Intelligent World”Your Trusted ICT Product Service Provider!
Sandy Yang/Global Strategy Director
WhatsApp / WeChat: +86 13426366826
Email: yangyd@qianxingdata.com
Website: www.qianxingdata.com/www.storagesserver.com
Business Focus:
ICT Product Distribution/System Integration & Services/Infrastructure Solutions
With 20+ years of IT distribution experience, we partner with leading global brands to deliver reliable products and professional services.
“Using Technology to Build an Intelligent World”Your Trusted ICT Product Service Provider!