CEO Interview with Adi Gelvan of Speedata

Key Takeaways

• APU achieves up to 100× performance boost over CPUs and GPUs
• Pharma deployment cut analytics runtime from 90 hours to 8 hours
• Customer replaced 38 servers with 3, slashing infrastructure costs >90%
• Speedata integrates with sovereign cloud Nebul for data‑jurisdiction compliance
• Zero‑code integration lets existing Spark jobs run on APU instantly

Pulse Analysis

The analytics bottleneck has become a strategic choke point for enterprises that rely on massive SQL‑driven ETL pipelines and AI data‑preparation. General‑purpose CPUs struggle with the deterministic, join‑heavy patterns of Spark SQL, while GPUs excel at unstructured floating‑point math but incur overhead for relational workloads. Purpose‑built silicon, exemplified by Speedata’s APU, addresses this mismatch by embedding relational operators directly into the processor’s datapath, delivering order‑of‑magnitude speedups without sacrificing existing software ecosystems.

Real‑world deployments illustrate the economic impact. In a pharmaceutical data‑pipeline, the APU trimmed a 90‑hour batch job to just eight hours, a 275× acceleration that translates into faster drug‑discovery cycles. A global tech firm consolidated 38 Spark servers down to three, cutting infrastructure spend by more than 90% and dramatically reducing energy consumption. These gains are not limited to batch jobs; the APU also speeds AI data‑preprocessing, enabling larger training datasets and quicker model iteration, a critical advantage as enterprises embed large language models into business processes.

Speedata’s competitive edge lies in its frictionless integration and focus on emerging use cases like agentic analytics, where AI agents generate SQL queries on‑the‑fly. Unlike CPUs, ARM chips, or repurposed GPUs, the APU is architected for deterministic relational workloads, offering predictable latency and lower total cost of ownership. The partnership with Nebul’s sovereign cloud further differentiates Speedata by meeting stringent data‑jurisdiction regulations, a growing concern for European firms. As AI‑driven analytics scale, the demand for such specialized compute is poised to reshape the data‑center landscape, positioning Speedata as a potential disruptor in the high‑performance analytics market.