Tel Aviv-based NeuroBlade, which offers data analytics tools that claim to use processing-in-memory, raises $83M Series B, bringing total funding to $110M
Meir Orbach / CTech :
Context & Ripple Effects
Two years after its $23M Series A backed by Marius Nacht and Intel Capital, NeuroBlade has quadrupled its raise to $83M — $110M total — for processing-in-memory silicon aimed at data analytics workloads. The round lands a week after rival Speedata came out of stealth with $70M across seed and Series A for its own custom analytics chips.
Together the two rounds put Tel Aviv at the center of a narrow but crowded bet: that moving compute into memory can beat general-purpose processors on analytics cost and speed. Weka's $135M Series D on the data-management side shows the surrounding stack is raising at the same cadence.
First-order effects
- NeuroBlade gets the capital to move its processing-in-memory chip from claim to shipping product, with Intel Capital already inside the cap table as a strategic signal to server buyers.
Second-order effects
- Speedata now competes head-to-head for the same customers and talent, pushing both vendors to land design wins quickly or concede that analytics silicon is a one-winner market per architecture.
Third-order effects
- If the pattern holds, hyperscale analytics stacks restructure around purpose-built memory-compute hardware rather than CPU upgrades — with Israeli chip startups, not US incumbents, supplying the specialized layers.
The trend: Custom data-analytics silicon is drawing nine-figure backing out of Tel Aviv, positioning Israeli chip startups as the suppliers of the post-CPU analytics layer.