Mobilint
Category: AI Chips / Semiconductors
A fabless startup developing high-performance Neural Processing Units (NPUs) and AI SoC solutions designed to bridge the gap between heavy AI workloads and power-constrained edge/on-premise environments. Mobilint was founded in 2019. The company is led by Dongjoo Shin. Based in Seoul, South Korea. Team size: 101-500. Total funding raised: $61.7M (KRW 100B+). Latest round: Series C. Key investors include Intervest, KDB Bank, Korea Investment Partners, LB Investment, FuturePlay, Kyobo Securities, Union Investment Partners, Daesung Private Equity, L&S Venture Capital, Praxis Capital Partners, POSCO Investment, Company K Partners.
- Founded
- 2019
- Headquarters
- Seoul, South Korea
- Team size
- 101-500
- Total funding
- $61.7M (KRW 100B+)
Value proposition
Delivers a superior balance of high-computational performance (up to 80 TOPS) and extreme energy efficiency compared to traditional GPUs, supported by a developer-friendly SDK that ensures 99.9% model accuracy after optimization.
Products and solutions
ARIES: High-performance NPU chip (80 TOPS) for on-premise AI servers and industrial applications., MLA100: ARIES-based PCIe accelerator card for high-intensity AI inference tasks., MLA100 MXM: 80 TOPS NPU module in MXM form factor for embedded AI PCs (launched 2025)., MLA400: Quad-ARIES PCIe Gen5×16 accelerator card delivering 320 TOPS at 135W for multi-stream video analytics and on-prem LLM inference., MLX-A1: Independent AI Box (Mini PC) powered by ARIES NPU for standalone edge computing (CES 2026 Innovation Award)., REGULUS: Standalone low-power AI SoC (~10 TOPS, <3W) integrating CPU and NPU for on-device applications; commercial supply accelerating as of Aug 2026., qb SDK: Proprietary full-stack SDK/compiler; validated with 490+ models across vision, transformer, and multimodal architectures.
Unique value
Mobilint is a founding member of MLCommons (MLPerf), consistently outperforming global giants in AI benchmark tests for power-to-performance ratios in edge environments.
Target customer
Manufacturers of smart factory equipment, robotics and drone developers, CCTV and surveillance system providers, smart city infrastructure firms, and industrial PC/embedded system manufacturers.
Industries served
Smart Manufacturing & Factory Automation, Robotics & Autonomous Drones, Public Safety & Intelligent Surveillance (CCTV), Smart City & Traffic Infrastructure, Healthcare (Medical Imaging), Industrial & Embedded Computing
Technology advantage
Proprietary NPU architecture optimized for multi-channel video stream processing and edge-LLM tasks; 'qb' SDK enables seamless deployment of 490+ commercial AI models with near-zero accuracy loss; MLA400 scales to 320 TOPS via Quad-ARIES for on-prem multi-stream and LLM workloads.
How they differentiate
Optimized balance of high compute performance (up to 80 TOPS) and extreme energy efficiency for edge environments, supported by the proprietary 'qb' SDK that ensures 99.9% model accuracy retention after optimization.
Main competitors
FuriosaAI, Rebellions, Sapeon, Hailo, Graphcore
Key partnerships
Samsung Foundry (14nm FinFET manufacturing partner), SemiFive (SoC design house collaboration), Amazon Web Services (AWS) (Collaboration on NPU integration with AWS IoT Greengrass), Siemens Healthineers (Hardware/Solution collaboration), Kolon Benit (AI Alliance partner for enterprise distribution), Hyvision System (Smart factory and industrial solution 실증/PoC partner), MLCommons (Global benchmark body founding member), POSCO DX (Strategic partnership for NPU integration into PosMaster industrial control system, April 2026), Lotte Innovate (MOU for NPU deployment in smart infrastructure and Physical AI, April 2026), INTOPS (MOU for joint AI manufacturing solutions development, March 2026), Nota AI (AI optimization technology supply and strategic partnership, May 2026), DFI (Computex 2026 MOU for industrial AI systems with Mobilint NPU, June 2026)
Notable customers
Samsung Foundry (Manufacturing Partner), Amazon Web Services (AWS IoT Greengrass Collaboration), Kolon Benit (Distribution Partner), Siemens Healthineers, Hyvision System, POSCO DX (Strategic customer for NPU-powered smart factory systems), Lotte Innovate (Pilot deployment of NPU in industrial sites), IntelliVIX / Korea Expressway Corporation (MLA400 AI servers for ~4,000 highway CCTV feeds, Aug 2026), Republic of Korea Navy (NPU AI video analytics on 50 ships, Aug 2026)
Major milestones
Founding member of MLCommons (MLPerf) in 2020, Successfully secured Series B funding of 20B KRW in Jan 2024, Won CES 2025 Innovation Award for 'REGULUS' (Standalone AI SoC), Commenced mass production of ARIES NPU and MLA100 accelerator in January 2025, Won CES 2026 Innovation Award for MLX-A1 Edge AI Box, Closed ₩70B Series C funding round in April 2026, Signed strategic partnership with POSCO DX for NPU-powered smart factories, Signed MOU with Lotte Innovate for smart infrastructure NPU deployment, Signed NPU integration MOUs with Taiwanese firms DFI, NEXCOBOT, and TOP MEASURE at Computex 2026 (June 2026), Accelerating REGULUS AI SoC commercial supply in Korea and global markets (August 2026), Deploying MLA400-based AI servers with IntelliVIX for ~4,000 Korea Expressway CCTV cameras (August 2026), Selected to supply NPUs for AI video analytics on 50 ROK Navy ships (August 2026)
Growth metrics
Cumulative private funding ~₩100B (~$61.7M) incl. ₩70B Series C (Apr 2026); ARIES in mass production; REGULUS AI SoC commercial supply accelerating domestically and globally as of Aug 2026; headcount ~115 (Jul 2026); major public deployments include Korea Expressway ~4,000 CCTV and ROK Navy 50-ship AI video pilot.
Market positioning
Edge AI NPU challenger focusing on industrial IoT, robotics, and on-premise AI acceleration rather than large-scale cloud data centers.
Geographic focus
South Korea, North America, and Global Industrial Markets
Patents and IP
Holds a robust domestic and international patent portfolio focused on NPU architecture, energy-efficient ASIC design, and proprietary deep learning acceleration algorithms.
About Dongjoo Shin
Dongjoo Shin is a prominent expert in AI semiconductor design with over a decade of research experience. He holds a PhD from KAIST, where his work focused on high-performance Neural Processing Units (NPUs). He is credited with developing the world's first Edge ASIC chip for artificial intelligence and currently serves as a member of the Presidential Advisory Council on Science and Technology in South Korea.
Latest news about Mobilint
Also named in
More AI Chips / Semiconductors companies
Official website: https://www.mobilint.com