What Happened Recent announcements span three layers of the AI stack. NVIDIA is making Run:ai open source for GPU workload orchestration, while promoting CUDA-X libraries for semiconductor simulation and cuDF…

At GTC Berlin, banks, payment companies and financial data providers are showing what production-scale intelligence looks like — built on NVIDIA accelerated computing, powered by open weight models and running on their own infrastructure. Revolut is rebuilding its technology stack on a single foundation model PRAGMA, powered by NVIDIA accelerated computing. Using the NVIDIA cuDF […]

At GTC Berlin, banks, payment companies and financial data providers are showing what production-scale intelligence looks like — built on NVIDIA accelerated computing, powered by open weight models and running on their own infrastructure. Revolut is rebuilding its technology stack on a single foundation model PRAGMA, powered by NVIDIA accelerated computing. Using the NVIDIA cuDF […]

The cuDF skills guide testing, conversion, benchmarking, and optimization for rewriting Spark UDFs into GPU-accelerated code.

* Benchmark analysis shows years of LSEG Tick History can be analysed in seconds using NVIDIA cuDF for GPU-accelerated data processing, reducing processing time from hours to seconds. * Enables firms to bring deeper historical analysis closer to decision-making as markets move towards near 24-hour trading LSEG today announced a collaboration with NVIDIA to help financial institutions analyse large volumes of LSEG Tick History data at significantly greater speed and scale using NVIDIA cuDF for accelerated computing on NVIDIA GPUs. This is achieved without creating another full copy of the underlying dataset while also preserving the licences, permissions and controls that govern access. The collaboration brings LSEG’s trusted, permissioned historical market data closer to the computing environments used for quantitative research, market analysis and AI. As part of the LSEG Everywhere AI strategy, users can run large-scale data processing on Tick History datasets to turn LSEG content into insights faster. Benchmark analysis shows years of LSEG Tick History can be analysed in seconds using NVIDIA cuDF and Polars, reducing processing time from hours to seconds. As equity trading moves towards near 24-hour operation, firms are moving away from legacy models that rely on market-close data loading and overnight analytics. This capability enables deeper, continuous historical analysis throughout the trading day. LSEG Tick History holds 110 trillion rows of data across 100 million instruments spanning more than 30 years and more than 580 exchanges, trading venues and contributors. In H1 2026, traffic over LSEG’s real-time network grew by 29% versus H1 2025, reaching a new peak of 26 million messages per second in June, while customer API requests for Tick History data increased by 48%. Ron Lefferts, Co-Head of Data & Analytics, LSEG, said: “This collaboration expands how customers can draw insight from LSEG’s trusted, governed Tick History. Customers want to work with detailed historical data at speed, without the burden and risk of creating another full copy. This helps firms analyse deeper histories closer to the point of decision, while keeping the permissions and controls they rely on fully intact.” Ioana Boier, NVIDIA, said: “AI is opening new ways to understand market behavior across vast amounts of data. NVIDIA cuDF helps researchers analyze LSEG Tick History at scale without downsampling away the detail that makes the data valuable for advanced analytics and AI.” The collaboration is intended to support customer workflows including: * Quantitative research and strategy testing: examining longer histories, additional trading venues and a wider range of market conditions without reducing the detail of the underlying data. * Execution and market analysis: reconstructing order books and historical market conditions, and analysing pricing, liquidity and execution quality at greater depth. * Risk, market replay and AI-enabled research: accelerating scenario analysis and historical market replay, and creating a foundation for customers to apply their own analytical and AI models to governed LSEG data.

AIエージェントでSpark UDFをGPU高速化。

Learn how GFQL indexes speed up Cypher graph queries on pandas, Polars, and cuDF dataframes, with measured lookup speedups of up to 518x.

In this tutorial, we implement NVIDIA cuML as a GPU-accelerated machine learning framework and build a practical workflow that demonstrates how RAPIDS can accelerate familiar data science and machine learning tasks. We begin by configuring the GPU environment and examining cuml.accel, which lets us accelerate existing scikit-learn workloads with minimal code changes, before moving to the native cuML API for direct CuPy and cuDF interoperability. We then benchmark CPU and GPU implementations of PCA, K-Means, nearest-neighbor search, logistic regression, random forests, and DBSCAN, while using synchronized timing to obtain meaningful performance measurements. We also build GPU-based manifold-learning and clustering pipelines with UMAP, t-SNE, HDBSCAN, and trustworthiness metrics; explore high-throughput forest inference with FIL; validate GPU-generated SHAP explanations; perform hyperparameter optimization with scikit-learn meta-estimators; and finally serialize trained models while examining portability between GPU and CPU environments. We configure the tutorial environment, define dataset sizes and benchmarking utilities, and verify that an NVIDIA GPU is available. We install and initialize RAPIDS cuML when necessary, set up CuPy and reproducibility controls, and create synchronized timing and result-tracking helpers. We also demonstrate cuml.accel by running an unmodified scikit-learn workload and comparing its CPU execution with GPU-accelerated execution. We work directly with the native...

OneRail has launched an AI-powered delivery platform that uses Nvidia technology to help retailers, wholesalers, and distributors decide how individual orders should be delivered. Called OmniSTAR, the system evaluates options including owned fleets, couriers, parcel carriers, and other delivery modes, then selects the lowest-cost option that meets the required service level, according to OneRail. The platform combines Nvidia’s cuOpt decision optimisation engine and cuDF data processing software with OneRail’s delivery pricing and performance data. Nvidia accelerated computing infrastructure is used to process the routing and delivery-mode calculations. OneRail said the system can reduce computation times by as much as 10 times.

Cloudera is embedding NVIDIA's cuDF library into its Data Engineering platform to deliver up to 4x Spark acceleration without code changes.

Native NVIDIA cuDF integration for Apache Spark 4.1 targets data preparation time and cloud compute costs while extending acceleration across Cloudera’s hybrid platform.

ClouderaはNVIDIAと提携し、Cloudera Data EngineeringにおいてApache Spark 4.1向けNVIDIA CUDA-X cuDFアクセラレーションのネイティブサポートを発表した。既存のPySparkやSQLアプリケーションを書き直すことなく、エンタープライズ企業がGPUを直接活用できるようになる。これにより、Spark処理が最大4倍高速化され、ハイブリッドおよびマルチクラウド環境におけるAI用データ準備のコストと時間が大幅に削減される。

YouTube video by Sky News Australia

Exploring GPU acceleration with cuDF, cudf.pandas, and the Polars GPU Engine

YouTube video by Thom Hartmann Program

YouTube video by Thom Hartmann Program

GPU powered dataframes in R. Contribute to bbtheo/cuplyr development by creating an account on GitHub.