Silicon IP and edge AI software for wireless and sensing products
Ceva licenses wireless communications and edge AI IP to semiconductor and device makers. The stack reveals a hardware-software hybrid: DSP cores, 5G/Wi-Fi/Bluetooth/Ultra-Wideband radio IP alongside PyTorch and TensorFlow for inference, plus graph compiler tools (MLIR, custom stack) for NPU workloads. Active projects span cellular modems, radar/lidar baseband, and AI graph compilation — signaling a pivot from pure connectivity IP toward inference-at-edge as a co-equal business pillar. Engineering-dominated hiring (18 of 20 roles) skews senior/lead, indicating depth in specialized silicon and compiler work rather than rapid headcount scaling.
Notable leadership hires: Business Operations Director
Ceva develops and licenses IP blocks for wireless connectivity (5G NR, Wi-Fi, Bluetooth, Ultra-Wideband) and edge inference (DSP cores, deep learning accelerators, graph compilers) to OEMs and SoC designers in consumer IoT, mobile, automotive, and industrial segments. Founded in 2002 and publicly traded, the company operates design centers across Israel, Ireland, France, UK, and the US, with ~400 employees. Revenue derives from IP licensing to chipmakers; customers integrate Ceva cores into their own silicon for faster time-to-market and lower development risk. Current focus includes next-generation modems, radar/lidar processing, and AI compilation tools for NPU-based systems.
C++, Python, SystemVerilog, PyTorch, TensorFlow, 5G NR, Wi-Fi 802.11, Bluetooth, Ultra-Wideband, Cadence, FPGA/ASIC tools, Docker, Kubernetes, LLVM, MLIR, and GitLab CI/CD. Recent focus on graph compiler infrastructure (MLIR-based) for AI inference.
AI graph compiler software for NPU systems, next-generation connectivity IP (cellular modems, radar/lidar baseband), Ceva Sensing SDK, graph compilation flows for edge AI workloads, and macOS driver/tools development.
Other companies in the same industry, closest in size
Ceva, Inc.'s technology stack, projects, and hiring signals are inferred from public hiring and company data — career pages, public listings, and company web presence — then clustered and de-duplicated. Figures are estimates that refresh over time. Read our full methodology →
This is not an official vendor or customer list. It is a technology-adoption signal inferred from public data, intended for B2B research.