OhpiNeuro
Category: AI Chips / Semiconductors
South Korean startup developing the N260 series of ultra-low-power neuromorphic AI processors (SNN-based) for edge AI devices, plus DX-MRAM-based next-gen AI semiconductors. OhpiNeuro was founded in 2025. The company is led by Jeon Byeong-du (전병두). Based in Giheung-gu, Yongin, Gyeonggi-do, South Korea. Team size: 1-10. Latest round: Bootstrapped.
- Founded
- 2025
- Headquarters
- Giheung-gu, Yongin, Gyeonggi-do, South Korea
- Team size
- 1-10
Value proposition
Develops next-generation AI semiconductors that solve the memory bottleneck and high power consumption of conventional AI processors by applying Spiking Neural Network (SNN)-based neuromorphic computing that only computes when data changes, dramatically improving energy efficiency for edge AI devices.
Products and solutions
N260 series neuromorphic AI processors: N260 Lite (≤0.7W max power, ≤28nm process, ≤5×5mm chip, targets smart cameras, industrial IoT, intelligent sensors, drones, mobile robots, development completion targeted H2 next year) and N260 Agent (next-gen processor, development starts H1 next year), DX-MRAM (DX-TR based MRAM) next-gen AI semiconductor platform for AI agents and large on-device AI models.
Unique value
Ultra-low-power neuromorphic AI processors (N260 series) using SNN computing that only activates on data change, targeting ≤0.7W power consumption on ≤28nm process in ≤5×5mm form factor, plus a DX-MRAM-based next-gen AI platform for on-device AI and AI agents.
Target customer
Edge AI device makers across smart cameras, industrial IoT, intelligent sensors, drones, and mobile robots; global edge AI market
Industries served
Edge AI / on-device AI, smart cameras, industrial IoT, intelligent sensors, drones, mobile robotics, AI agents
Technology advantage
SNN (Spiking Neural Network)-based neuromorphic computing that only performs computation when data changes, maximizing energy efficiency; DX-MRAM (DX-TR based MRAM) memory technology to overcome MRAM limits, improving memory density and energy efficiency for next-gen AI processors.
How they differentiate
Differentiates via ultra-low-power SNN-based neuromorphic architecture (≤0.7W) combined with proprietary DX-MRAM (DX-TR based MRAM) memory technology, targeting the edge AI market where power efficiency and memory innovation are the key competitive levers.
Main competitors
NanoLatis (나노라티스, Korean neuromorphic semiconductor startup, also DIPS-selected, seed from Korea Investment Accelerator), OpenEdge Technology (오픈엣지테크놀로지, Korean NPU IP company), international neuromorphic players (IBM TrueNorth, Intel Loihi, BrainChip)
Key partnerships
Selected for 2026 Ministry of SMEs and Startups 'Super Gap Startup Project (DIPS)' government support program (run with Seoul National University System Semiconductor Industry Promotion Center), DIPS provides up to ~2억원/year for 3 years for tech development and commercialization.
Major milestones
Founded 2025-08-01, selected for 2026 Ministry of SMEs and Startups 'Super Gap Startup Project (DIPS)' government program, developing N260 Lite (target H2 next year completion) and N260 Agent (H1 next year start).
Market positioning
Early-stage Korean neuromorphic AI semiconductor startup selected for the government's 2026 'Super Gap Startup Project (DIPS)' program, positioning to enter the global edge AI market.
Geographic focus
South Korea (domestic) with global edge AI market ambitions
About Jeon Byeong-du (전병두)
Ex-CoreAce (코아에이스). Background in semiconductor/AI industry.