Israel-based Speedata, which is developing custom chips for data analytics processing, emerges from stealth with $55M Series A and $15M seed
Context & Ripple Effects
Speedata's stealth exit lands in the middle of an Israeli cluster of startups attacking data-center silicon from different angles: NeuroBlade raised an $83M Series B for processing-in-memory analytics within a week of this round, and Hailo had just closed a $136M Series C at a ~$1B valuation for edge AI chips.
The arc since: Pliops, which accelerates databases on flash, later raised a $100M Series D, and Speedata itself needed until 2025 and a $44M Series B before showing its first analytics processing unit — evidence of how long and capital-hungry the road from this kind of Series A to shipping silicon is.
First-order effects
- Speedata exits stealth with roughly $70M across seed and Series A, giving it the war chest to move its custom analytics chip from design toward silicon while competing head-to-head with NeuroBlade's processing-in-memory approach for the same big-data workloads.
Second-order effects
- Data-center investors now face a fork between rival analytics architectures — Speedata's APU versus NeuroBlade's in-memory processing versus Pliops's flash-side acceleration — forcing them to pick a horse in a category where each bet requires nine-figure follow-on rounds.
Third-order effects
- If the funding cadence holds, analytics processing becomes a distinct accelerator category alongside AI chips, with Israel as its densest hub — but Speedata's four-year gap between this round and its first APU suggests consolidation pressure on startups that can't sustain multi-round raises before revenue.
The trend: Specialized analytics processors are emerging as their own silicon category, concentrated in Israel and funded on the same multi-hundred-million-dollar, multi-year path as AI accelerators.