Mobilint MLA100 obtains 'TTA Verified' first certification…validates reliability of industrial and public AI infrastructure
Published2026.09.22 16:51
System compatibility, operational stability, and AI model execution environment verification completed
AI semiconductor company Mobilint has obtained official third-party verification of the reliability of its self-developed neural processing unit (NPU)-based AI accelerator. It is expected to facilitate the expansion of edge and on-premises AI infrastructure deployment in industrial and public sectors.
Mobilint announced that its AI accelerator MLA100, equipped with the self-developed NPU 'ARIES,' has obtained 'AI Accelerator TTA Verified' certification from the Korea Information and Communications Technology Association (TTA). MLA100 is the first AI accelerator product to receive this certification.
This certification is characterized by comprehensive verification not only of computational performance, but also of system compatibility, operational stability, software environment, and AI model execution capability—all essential for actual field deployment.
MLA100 is a PCIe-based AI accelerator equipped with ARIES, Mobilint's proprietary NPU. It is designed to be installed on existing industrial servers or edge devices to perform AI inference tasks such as video analysis, vision inspection, and equipment anomaly detection.
The product delivers maximum AI computational performance of 80TOPS with power consumption of approximately 25W. It also adopts a standard PCIe interface to ensure interoperability with existing systems.
The TTA certification program evaluates hardware stability, PCIe compatibility, server and platform interoperability, operational stability under varying environment conditions, as well as SDK and software environment, and actual AI model execution performance.
Mobilint conducted evaluations over approximately three months beginning in June of this year and obtained final certification in August. The evaluation results confirmed that MLA100 operates stably in various system environments and meets the conditions necessary for AI model execution and system integration.
Based on this certification, the company plans to expand application scope to fields requiring real-time AI inference such as video surveillance, smart manufacturing, public safety, and defense and aerospace.
Comments
Related news
Mobilint Unveils NPU-Based Security AI Solution at ISEC 2026
AI semiconductor specialist Mobilint will participate in ISEC 2026, Korea's largest information security exhibition held at Seoul COEX on August 11-12, and unveil NPU-based security AI solutions. Leveraging its AI semiconductors including MLA100, MLA400, MLX-A1, and REGULUS, the company will demonstrate real-time video analysis, AI security monitoring, and generative AI-based operational automation, presenting solutions that simultaneously meet high inference performance, data security, and power efficiency in edge environments.
2026.07.29by 배종인 기자
Mobilint officially launches AI USB 'MLD-R1'
Mobilint has officially launched the AI USB accelerator 'MLD-R1'. Designed based on the REGULUS SoC, this product can add on-device AI inference capabilities to notebooks, industrial PCs, and embedded systems with just a USB connection. Supporting a maximum performance of 10TOPS with approximately 3W low-power design, it enables operation of 1~3B scale LLMs, facilitating rapid AI adoption in industrial sectors such as smart factories, robotics, and medical devices.
2026.08.07by 배종인 기자
Domestic AI semiconductors, moving beyond demonstration into everyday use...Can they shake 'Nvidia's dominance'?
Domestic AI semiconductors are advancing beyond the research and pilot stage and being deployed in real environments including public services, industrial sites, and cloud platforms, with commercialization competition now entering full swing. Rebellio and FuriosaAI are expanding by securing domestic and international references with data center NPUs, while Mobiilinc is broadening public procurement and industrial applications centered on edge and on-device markets. However, Nvidia's core competitiveness lies not in the chip itself but in its vast software and developer ecosystem; for domestic NPUs to succeed, rapid model support, transparent performance disclosure, and complete product supply based on global partnerships are essential. Over the next two to three years, market trust and ecosystem development are expected to determine survival prospects.
2026.08.07by 배종인 기자
Mobilint and Intellibix Participate in Highway AI CCTV Conversion Project
Mobilint and Intellibix are participating in Korea Expressway Corporation's public AI CCTV conversion project, converting approximately 4,000 CCTVs across the nation's expressways into intelligent control systems. Based on the MLA400 AI accelerator, they are building an inference infrastructure that combines Vision AI and VLM to automatically detect emergency situations such as traffic accidents, stopped vehicles, and wrong-way driving. Phased AI adoption is possible through integration with existing video management equipment, and this represents an expansion of domestic AI semiconductor applications in public infrastructure.
2026.08.11by 배종인 기자
NOTA Selected as Lead Organization for K-On-Device Humanoid National Strategic Project
NOTA has been selected as the lead organization in the humanoid sector for the Ministry of Trade, Industry and Energy's national strategic project 'K-On-Device AI Semiconductor Technology Development Project'. In collaboration with LG Electronics and Mobilint, NOTA will lighten the vision language action model (VLA) and convert it to match the domestic NPU architecture, enabling real-time on-device inference within robots. Through optimization techniques including token compression, quantization, and pruning, the company will reduce cloud dependency and minimize latency while securing real-time autonomous control performance.
2026.09.08by 배종인 기자
