speedata.ioCrawled Nov 7, 2026
The Speedata APU processes Apache Spark and AI workloads while keeping data in AWS, Azure, or GCP, with up to $10,000 in free APU hours available.
Dramatic hardware acceleration for analytics workloads
The page offers a test to compare the speed of Spark workloads on Speedata’s APU, using CLI, browser uploads, or TPC-DS benchmarks.
right at the hardware layer, with zero code changes
Microsoft's CoddSpeed paper won the Best Industry Paper Award at SIGMOD 2026. Inside it: the Speedata Analytics Processing Unit (APU), integrated and benchmarked by Microsoft, delivering up to 259× speedup on TPC-H queries.
The page examines designing a domain-specific processor for analytics by building silicon rather than optimizing algorithms for CPUs.
Workload Analyzer - Collecting Spark logs
The Speedata APU achieves high throughput by mapping processing into its hardware pipeline, enabling parallel row processing across hundreds of steps.
Dan is Chief Revenue Officer at Speedata, overseeing go-to-market strategy, revenue growth, and partnerships. He previously drove sales at TigerGraph and expanded the datacenter market at Xilinx.
Dror leads Project Management at Speedata, overseeing execution across the broader organization, with a focus on accelerating delivery and cross-team alignment.
Gal leads software engineering at Speedata, with over 20 years of experience in data and analytics, specializing in large-scale systems for major companies.
Liraz is Chief Administrative Officer at Speedata, managing business, human resources, and operational functions. She previously coordinated programs at Technion and worked in electrical engineering.
Speedata introduces the first purpose-built Analytics Processing Unit (APU) for accelerating database analytics and AI workloads, combining hardware and software into a unified platform.
SPEEDATA-Analytics at the Speed of Silicon Accelerate Apache Spark by 100x - right at the hardware layer, with zero code changes