Fraud Detection
IRS investigators needed to search multi-terabyte tax and transaction data to expose fraud and identity-theft rings, but CPU clusters and manual data stitching could not finish graph-heavy jobs quickly enough to support mission-critical investigations.
Reported impact
The IRS achieved more than 20x speed improvements at half the cost for data engineering and data science workflows, moved from failed overnight CPU jobs to production-scale fraud analytics, and created a foundation it can reuse for future NLP and inference workloads in government operations.
The IRS modernized fraud analytics with Cloudera and NVIDIA GPU acceleration so graph analysis, machine learning, and Spark-based data engineering could run against very large tax datasets. Investigators start with decades of records and multi-hop relationship data, then push those workloads through Cloudera on NVIDIA GPUs to accelerate both preparation and analysis. Initial runs saw immediate speedups without code changes, and after NVIDIA data scientists reviewed the code paths, optimized Spark-GPU workflows delivered much larger gains at lower cost. The accelerated stack also opened the door for the IRS to extend beyond historical fraud analysis into faster AI inference and natural-language analytics, all on the same GPU-backed foundation. Instead of breaking the dataset apart across CPU servers and manually stitching results together, the agency can now evaluate much more of the full signal in one governed analytics environment.
Large tax and transaction datasets flow into Cloudera-based Spark and graph workloads, RAPIDS and GPU acceleration speed the data engineering and analytics stages, and investigators use the results to trace suspicious relationships and flag potential fraud or identity-theft patterns.
To combat tax fraud and uncover bad actors, IRS investigators must analyze decades' worth of data.
We're currently implementing this integration, and are already seeing over 20x speed improvements at half the cost.
With Cloudera running on NVIDIA GPUs, workloads immediately ran up to 5X faster with no code changes.
With improved computing infrastructure and AI implementation, the IRS is cutting costs and better protecting taxpayers by preventing fraud and identity theft.
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
NVIDIA Sample Jul 7, 2026 French Ministry of Economy and Finance French public servants needed to answer citizen requests against more than 100 million documents per year while maintaining legal accuracy, security, confidentiality, and compliance, but manual retrieval and analysis of… 1,000 work hours saved per month for a typical 100-user deployment DeepBrain · NVIDIA AI Enterprise · NVIDIA DGX platform NVIDIA Sample Aug 14, 2026 Shopee Shopee needed large language models that could handle multilingual ecommerce tasks across Southeast Asia, Taiwan, and Brazil, but general-purpose frontier models degraded sharply on regional languages, shopping… 113x growth from 3 billion to 340 billion monthly tokens Compass v3.5 · NVIDIA A100 GPUs · NVIDIA H100 GPUs NVIDIA Sample Jun 2, 2026 Revolut Revolut needed to replace siloed fraud, credit, marketing, and recommendation models that depended on months of feature engineering with a single behavioral foundation model that could learn from raw financial event… 16.1% improvement in credit risk prediction performance from NVIDIA Llama-Nemotron-Embed-1B-v2 NVIDIA H100 GPUs · NVIDIA Llama-Nemotron-Embed-1B-v2 · Nebius AI Cloud 