딥엑스, 포스코DX와 AI 반도체 기반 공장 자동화 및 지능형 물류 시스템 상용화 협력

  • 딥엑스 AI 반도체를 활용하여 공장 자동화 및 지능형 물류 시스템에 활용할 수 있는 대단위 AI 연산 처리기반 AI 솔루션 공동 개발
  • 사전 검증 과정에서 딥엑스의 컴파일러가 포함된 소프트웨어 ‘DXNN’을 배포해 딥엑스 NPU의 최신 알고리즘 지원, 우수한 AI 정확도, 높은 연산 효율 확인
  • 딥엑스의 AI 반도체가 고성능과 고효율을 앞세워 산업현장에서 수만대의 카메라를 동시 AI 연산 처리하고 자동화할 수 있는 혁신적인 솔루션이 될 것”

딥엑스가 AI 반도체 기반 공장 자동화 및 지능화 물류 시스템의 상용화를 위해 포스코DX와 기술 개발을 협력한다. 딥엑스는 4월 6일 목요일에 김녹원 대표이사, 포스코DX 기술연구소 윤일용 상무 등 양사 관계자들이 참석한 가운데 포스코DX와 AI 반도체를 활용한 제조, 로봇, 물류, 안전 등 산업현장에 적용 가능한 AI 솔루션 기술의 상용화를 위해 업무협약을 체결했다고 밝혔다.

딥엑스는 포스코DX와 진행된 사전 기술 검증 과정에서 딥엑스의 컴파일러가 포함된 소프트웨어 개발 키트(SDK, Software Development Kit)인 ‘DXNN’을 배포해 자사의 하드웨어와 소프트웨어의 구동성을 검증한 후 딥엑스 기술의 우수함을 입증했다. 특히 전력 대비 인공지능 연산 성능 효율이 높아 산업 현장의 자동화를 위한 AI 솔루션으로 상용화하기에 적합한 것으로 평가받았다.

양사는 이번 협력을 통해 포스코DX의 공장 자동화 및 지능화 물류 설비제어시스템인 포스마스터(PosMaster-PLC)에 딥엑스의 AI 반도체 기술을 탑재해 산업현장 설비시스템의 운영을 자동화할 수 있는 대단위 AI 솔루션을 공동 개발할 계획이다. 이번 협약으로 양사는 스마트 팩토리 구동부에 사용될 AI 기능을 딥엑스의 AI SoC(System on Chip) 제품에 탑재하고 구동성 및 양산 가능성 등을 검증한 이후 적용범위를 확대해 나갈 계획이다.

 

이번 협약으로 딥엑스는 IPC(Industrial PC, 산업용 PC) 1대 정도의 AI 성능이 필요한 분야에 적용할 수 있는 ‘DX-M1(23TOPS)’ 제품으로 올해 포스코DX와 테스트를 진행하고 공장에 적용할 계획이며, 추가적으로 대규모 단위 AI 연산이 필요한 서버 솔루션인 ‘DX-H1(서버당 22POPS – 초당 2.2경회 인공지능 연산)’도 테스트를 진행할 예정이다. 해당 칩 솔루션은 다양한 산업 현장의 설비시스템에 M.2 모듈 및 PCIe 카드 형태로 설치해 수십만대의 카메라 및 센서를 통해 들어오는 데이터를 신속하게 분석할 수 있는 고성능 제품 라인이다.   

 

딥엑스는 모든 곳에 AI 컴퓨팅을 구현하기 위한 AI 반도체를 개발하는 회사로서, AI 기술을 모든 곳에 존재하는 다양한 전자기기에 고성능•저전력으로 구현할 수 있다. 현재 다양한 엣지 및 서버 AI 응용 분야에 맞춰 성능과 기능이 최적화된 AI 반도체 4종을 보유하고 있으며, AI 반도체 원천기술과 관련한 국내외 특허도 150여개 이상 확보한 상황이다. 국내에서는 대표적인 AI 반도체 팹리스로 평가받으며 글로벌 AI 반도체 시장을 목표로 한다.

 

김녹원 대표이사는 “포스코DX의 스마트 팩토리 기술과 딥엑스의 AI 반도체 기술을 결합하면 유럽형 4차 산업 혁명의 목표인 AI 기반 공장 자동화를 구축할 수 있어, 제조 산업 현장의 혁신을 선도할 수 있을 것으로 기대한다. 대규모 인공지능 연산 처리를 위해 딥엑스가 준비한 AI 반도체 기반 서버 솔루션인 DX-H1을 실증할 수 있는 기회를 얻게 되어 기쁘다”라며 “포스코DX와의 협업으로 포항제철의 2만대의 카메라 정보를 동시 처리하여 자동화 함으로써, 미래 스마트 팩토리 시장을 선도하는 데 중요한 기반을 구축할 수 있도록 지속적으로 협력할 것”이라고 말했다.

 

포스코DX 기술연구소 윤일용 상무는 “포스코DX의 인공지능 SW 및 제어 융합 기술과 딥러닝에 최적화된 인공지능 반도체의 연산 능력이 더해져 하나의 통합된 자동화/무인화 시스템을 구성할 수 있는 뜻깊은 협업 기회로 생각한다.”며 이를 통해 GPU에서 탈피한 더 고효율 연산 성능과 신뢰성 있는 솔루션을 제공하며 유연한 분산형 구조의 아키텍처를 제안함과 동시에 전력소비를 줄여 ESG경영에도 기여할 수 있을 것”이라고 전했다.

DEEPX Unveils Total AI Solutions for NVRs and Cameras at ISC West Elevates On-Device AI Ecosystem for Real-Time Video Analysis

AI Solutions for NVRs and Cameras at ISC West Elevates On-Device AI Ecosystem for Real-Time Video Analysis



Las Vegas, NV – [
March 31]DEEPX, a leading provider of high-performance, low-power AI accelerator solutions, is unveiling its latest innovations at ISC West 2025 in Las Vegas (March 31-April 4, 2025). The company is debuting a complete lineup of advanced AI products, including M.2 and PCIe-based acceleration solutions for Smart NVRs and edge servers, as well as the powerful V3 AI SoC designed for smart cameras and AI home hubs.

Built on DEEPX’s proprietary neural processing architecture, the new M.2 and PCIe AI accelerator cards deliver exceptional power efficiency and compute performance. These solutions enable OEMs and ODMs to reduce total cost of ownership (TCO) while dramatically enhancing real-time video analytics and AI inference capabilities. The compact M.2 cards, delivering up to 25 TOPS, are ideal for upgrading legacy NVRs with on-board AI. Meanwhile, the PCIe solutions offer over 100 TOPS of scalable acceleration for multi-stream, high-resolution video workloads in edge servers and gateways.

Also featured at this show is the DEEPX V3 AI SoC, purpose-built for smart cameras and home AI hubs. With 13 TOPS of AI performance, the V3 enables on-device vision intelligence such as facial recognition, person detection, and event-based alerts—all with ultra-low power consumption. Integrated H.265/H.264 encoding and decoding, along with an advanced ISP pipeline, make it a complete edge AI solution for next-generation smart home devices.

Why DEEPX Outperforms GPUs
DEEPX’s specialized AI architecture is designed for inference, delivering higher performance per watt than traditional GPUs. By focusing on matrix-based operations at the hardware level, DEEPX accelerators handle large-scale data in real time with significantly lower power and without cooling requirements—ideal for edge deployments with tight thermal and power constraints. Their scalable performance also helps OEMs and ODMs balance cost, size, and throughput. With on-device AI capabilities, DEEPX solutions further reduce latency and reliance on data centers, creating a more secure and efficient deployment environment.

Total AI Solutions for User-Friendly and Reliable Development
By partnering with Network Optics and CVEDIA—global leaders in video management solutions (VMS)—DEEPX delivers an all-encompassing AI solution that empowers physical security companies to quickly adopt advanced camera-based surveillance, AIoT frameworks, and edge AI capabilities. The seamless integration of DEEPX’s high-performance AI chip with Network Optics and CVEDIA’s VMS ensures real-time video analysis, anomaly detection, and object recognition, helping security enterprises efficiently detect and respond to threats. This collaborative effort not only reduces operational burdens and enhances situational awareness but also enables new revenue opportunities through scalable, user-friendly AI-driven services tailored to the evolving demands of modern security environments.

“We are witnessing explosive growth opportunities for edge devices in the physical security sector—such as AI-enabled cameras and IoT sensors,” said Mr. Jun, DEEPX’s Global Sales Lead. “We believe the on-device AI technology DEEPX is showcasing at ISC West will set a new standard for the market.” DEEPX invites ISC West attendees to visit Booth [33064] for live demonstrations of its full AI portfolio and to meet with technical and business leaders about integration opportunities.

DEEPX at Embedded World 2025: Showcasing Global Collaborations and Preparing for Mass Production March 11–14 | Nuremberg, Germany

DEEPX booth at Embedded World 2025 showcasing AI chip solutions

 

  • DEEPX showcases on-device AI solutions at Embedded World in Nuremberg, unveiling global collaborations.
  • The company plans to launch its first mass-produced AI semiconductor product by mid-year.

Nuremberg, Germany – DEEPX (CEO Lokwon Kim) will participate in Embedded World, held from March 11 to 14 in Nuremberg, Germany, to expand the on-device AI market and ecosystem in Europe by partnering with embedded system manufacturers and distributors. At this year’s exhibition, DEEPX will unveil the results of its collaborations with global players and emphasize its readiness to supply on-device AI solutions that meet worldwide market demand. During the event, DEEPX is presenting outcomes from joint projects at partner booths—including Micron, Raspberry Pi, AAEON, DFI, Portwell, SEEED, Biostar, Advantech, Inventec, Network Optics, and Embedded Artist. These live demos illustrate DEEPX’s leadership in on-device AI and its growing influence in industrial automation, robotics, physical security, smart manufacturing, and IoT.

Industrial and Commercial IoT Hardware and Solutions

DEEPX is collaborating with Raspberry Pi, along with Sixfab—a global provider of hardware for remote monitoring, control, and data collection. By also partnering with Revolution Pi and Edutech, DEEPX is delivering customized, AI-enabled solutions designed for enterprise IoT scenarios.

Driving AI, IoT, and Edge Computing with Micron
DEEPX showcases high-performance AI computing by integrating Micron’s LPDDR4 and LPDDR5X memory into multiple form factors (such as M.2 and PCIe). As AI, IoT, and edge computing sectors expand, DEEPX and Micron together address diverse industry needs by combining advanced memory and AI semiconductor technologies.

Factory Automation and Robotics with Global Industrial PC Leaders
DEEPX demonstrates industrial computing products powered by its DX-M1 M.2 module through partnerships with Inventec, AAEON, DFI, Portwell, SEEED, Biostar, Advantech, AWS Greengrass, and Onlogic. These AI-driven solutions enhance factory automation, medical systems, and smart factory operations. Portwell and DFI have publicly lauded DEEPX’s technology for delivering lower power consumption, reduced operating costs, and robust AI features—giving businesses remarkable efficiency, scalability, and stability.

Total AI Solutions for User-Friendly and Reliable Development
Working with Network Optics, a global leader in video management solutions (VMS), and Embedded Artist, which specializes in embedded system development and training platforms, DEEPX provides a complete AI solution for camera-based physical security, AIoT systems, and edge AI development environments. Notably, Network Optics’ VMS integrates seamlessly with DEEPX’s AI chip for real-time video analysis, anomaly detection, and object recognition. Embedded Artist further accelerates the adoption of DEEPX solutions, helping reduce time-to-market for IoT and AI projects across industries such as factory automation, logistics, and retail.

Aggressive Preparation for Mass Production
DEEPX is in the final stages of reliability testing and certification for its AI semiconductor chips and plans to launch its first mass-produced product in mid-2023. After conducting technology verifications with more than 300 global companies—spanning robotics, factory automation, physical security systems, and on-premise servers—DEEPX is now actively supporting the mass-production needs of over 20 clients.

DEEPX ramps up mass-production chip supply for the global market, bolstering tangible business expansion


AI semiconductor company DEEPX (CEO: Lokwon Kim) has been steadily increasing its global marketing efforts since introducing sample chips for international customer evaluation. Now, DEEPX is ramping up the supply of mass-production chips, paving the way for broader commercial opportunities across global markets.

Last month at CES 2025 – one of the world’s largest electronics trade shows – DEEPX was recognized as a “must-visit company” by the event’s organizer, the Consumer Technology Association (CTA), shining a spotlight on the global market. During the show, over 16,000 visitors stopped by the DEEPX booth, resulting in more than 100 requests for mass-production chip samples—strong indicators of the company’s expanding commercial prospects.

During CES, DEEPX showcased various applications of its chips in edge devices for smart cities, smart factories, and robotics, as well as in industrial equipment and edge servers for global enterprises. Demonstrations and on-site meetings were held with a wide range of customers and partners from the United States, Taiwan, Japan, and Europe. In addition, major Korean corporations, government officials, and investors visited the booth on tours to get a firsthand look at DEEPX’s technological strengths.

Among the highlights, DEEPX successfully ran its first-generation mass-production chips using AI algorithms developed by:

– Hyundai-Kia Motor Robotics Lab (robot platform)
– Inventec (smart factory)
– POSCO DX (smart factory & logistics automation)
– LG U+ (industrial and urban security)

This demonstration underscored the critical role of on-device AI semiconductors – beyond cloud-based approaches – in ushering in a new era of AI.

DEEPX also directly compared its solution to NVIDIA’s Jetson Orin, a GPU-based edge platform, demonstrating approximately 10 times the power efficiency and 20 times the price/performance efficiency of its own NPU-based product. By also conducting a ‘butter benchmark’ test against other leading global NPU technologies, DEEPX demonstrated its advantage in power efficiency, thermal management, and operational stability – critical factors for on-device AI.

To illustrate its broad compatibility, DEEPX showcased integrations with application processor (AP) systems-including NXP, Renesas, TI, Broadcom, Rockchip, and Intel-as well as small form factor computing solutions such as the popular Raspberry Pi. The company also demonstrated interoperability with partner industrial PCs and leading workstation and server systems from HP, Dell, Supermicro, Lenovo, Inventec, and others, highlighting the versatility and scalability of its solution.

Looking ahead, DEEPX will partner with LG U+ at MWC (Mobile World Congress) in Barcelona from March 3 to demonstrate on-device AI solutions powered by its first-generation chip. DEEPX will also host its own booth to showcase the roadmap for its DX-M2 on-device AI semiconductor, designed for large language models. Following MWC, DEEPX will head to Embedded World on March 11 in Nuremberg, Germany, where its products will be showcased at partner booths – including Micron, Raspberry Pi, AAEON, DFI, Portwell, SEEED, Biostar, Advantech, and Network Optix – as well as DEEPX’s own stand. The company will conduct additional live demos and meet with potential customers to explore new business opportunities.

News Source-(CTA) What Not to Miss at CES 2025

DEEPX Wins 2024 Product of the Year Award, Demonstrating Global AI Semiconductor Leadership


On-device AI semiconductor company DEEPX (CEO Lokwon Kim) announced on the 6th that it had received the 2024 Product of the Year award, hosted by Electronic Products, a publication under EE Times, a global electronics industry journal.

Now in its 49th year, this award is given based on criteria such as groundbreaking technological advancements, innovative design, superior price-performance ratio, and market potential. It is a prestigious accolade previously awarded to global semiconductor leaders such as NXP, Renesas, Marvell Technology, and Texas Instruments, marking DEEPX’s international recognition for its technological excellence.

DEEPX is leading the next-generation semiconductor market with its innovative AI technology and global competitiveness. The company has filed over 300 patents worldwide and registered over 70, securing world-class proprietary AI semiconductor technology assets.

Based on these technological achievements, DEEPX was awarded the Presidential Commendation at the 2023 Korea Intellectual Property Office’s ‘Invention Day’ event. The company was also recognized at Computex Taipei, a major global AI semiconductor event, winning an innovation award among 400 competing companies. Additionally, DEEPX won three CES 2024 Innovation Awards and became the first AI semiconductor company invited to the World Economic Forum’s ‘Davos Forum,’ where it led discussions on energy credit policies for AI computing.

Furthermore, DEEPX has been listed in the EE Times ‘Silicon 100’ for two consecutive years and was named the ‘Best Company in the AI Semiconductor Industry’ by U.S. market research firm Frost & Sullivan. DEEPX continues solidifying its position as a key player in the AI semiconductor market through its global achievements and recognition for its innovation and technological leadership.


News Source – Electronic Products

BIOSTAR showcases new partnership with DEEPX in conjunction with new x86-based Edge AI Solutions

x86 edge AI platform from Biostar and DeepX

 

BIOSTAR, a leading manufacturer of Edge AI embedded computers, IPC solutions motherboards, graphics cards, and PC peripherals, is excited to announce x86 Edge AI platform showcase at CES 2025. Scheduled to take place from January 7 to 10 in Las Vegas, Nevada, USA, BIOSTAR, in partnership with DEEPX, an AI semiconductor company from South Korea, will showcase cutting-edge x86 Edge AI solutions at DEEPX’s booth (#9045) in the North Hall of the Las Vegas Convention Center.

 

DEEPX is a pioneering company in on-device AI. It develops advanced AI semiconductors that optimize performance, reduce power consumption, and enhance cost efficiency across various industries, including smart camera modules, smart mobility, smart factories, consumer electronics, smart cities, surveillance systems, and AI servers. Building upon their successful collaborations, BIOSTAR and DEEPX set to unveil their latest joint innovations at CES 2025, further enhancing their prominence in the industrial computing ecosystem.

 

This strategic collaboration integrates BIOSTAR’s AMD B650 motherboards and storage device with DEEPX’s advanced AI accelerators—the DX-H1 PCIe x16 (Bifurcation x4x4x4x4) and DX-M1 (M.2) modules – operating on a Linux platform. The real-time demonstration will feature state-of-the-art object detection, highlighting the seamless integration and scalability of their combined technologies in Edge AI applications.

 

By integrating BIOSTAR’s x86 based edge AI platforms with DEEPX’s advanced AI capabilities, this partnership establishes a new benchmark for AI computing industry, highlighting BIOSTAR’s pivotal role in shaping the future of Edge AI technology. As DEEPX becomes synonymous with cutting-edge AI NPU technology, BIOSTAR’s collaboration with such an industry leader enhances its reputation as an innovative and reliable partner in the AI edge computing industry.

 

News Source-Tech Power up

 

Virtium Embedded Artists Introduces Integrated AI Accelerator DEEPX DX-M1 and iMX8M Mini Kit

DEEPX DX-M1 & i.MX8M Mini

 

Industrial and embedded computing specialist Virtium Embedded Artists has announced the launch of the DEEPX iMX8M Mini AI Kit, developed in partnership with DEEPX: a single-board computer (SBC) style development bundle with an “AI Booster” delivering 25 tera-operations per second (TOPS) of compute at a 5W power draw for on-device machine learning (ML) and artificial intelligence (AI) workloads.

 

“This product redefines the limits of edge AI performance, delivering unparalleled power and cost efficiency,” claims Virtium Embedded Artists managing director Anders Rosvall of the company’s newest launch. “The DEEPX iMX8M Mini AI Kit simplifies the evaluation process for design engineers, enabling them to determine if our solution meets their application requirements while accelerating time to market.”

 

The DEEPX iMX8M Mini AI Kit isn’t a true single-board computer, though it can be treated like one; rather, it’s three boards bundled together. The first is an iMX8M Mini uCOM computer-on-module, built around NXP Semiconductor’s i.MS 8M Mini processor — giving it four Arm Cortex-A53 cores running at up to 1.8GHz and a Cortex-M4 core running at up to 400MHz, a VeriSilicon Vivante GC Nano Ultra 2D graphics processor, and GC320 3D graphics processor. To this the company has added 1GB of LPDDR4 memory and 8GB of on-board eMMC storage.

 

The second board is DEEPX’s contribution: an M.2 module with the company’s DX-M1 AI Booster, delivering a claimed 25 tera-operations per second (TOPS) of minimum-precision compute at a 5W power draw — the highest cost efficiency, power efficiency, and performance efficiency in the world, the company boasts. The final board ties everything together: a carrier board for the uCOM module, which includes a gigabit Ethernet port, USB ports, microSD card slot, MIPI Display Serial Interface (DSI) and parallel RGB video outputs plus a bundled DSI-to-HDMI adapter, a MIPI Camera Serial Interface (CSI) input, mikroBUS header, expansion connectors, and a dedicated debug interface.

 

The kit, which can run pre-trained models or custom models compiled with a toolkit available from DEEPX, is now available to order for “qualified projects” on the Virtium Embedded Artists website; the company has not publicly disclosed pricing.

 

 

 

News Source (1) – ‘Energy-Efficient DEEPX iMX8M Mini AI Kit (Hacksters.io)

News Source (2) – Embedded Artists 

DEEPX with global leaders at CES 2025 showcasing AI semiconductor mass-production-ready achievements


DEEPX an on-device AI semiconductor company, announced that it will participate in CES 2025, the world’s largest electronics expo in Las Vegas, USA, next January to showcase its AI semiconductor mass-production ready achievements with global leaders.

At the exhibition, DEEPX will emphasize its core business strategy of “All-in All-On”. Recently, global companies have been declaring an ‘All-in’ strategy for the adoption of AI technology, and as AI commercialization is in full swing, an ‘All-on’ strategy for on-device AI technology that enables AI anytime, anywhere is also gaining attention. DEEPX’s “All-in All-on” strategy is a strategic step toward becoming a global leader in on-device AI semiconductors and serving as a catalyst for all of its customers to drive their AI strategies.

With this strategy, DEEPX is contributing to the reshaping of all camera-based systems, computing systems of all kinds, and autonomous mobile vehicle technologies based on AI solutions. To realize this, DEEPX has been developing low-power, high-performance AI semiconductor solutions. By realizing high-performance AI computation on-device, it delivers superior performance in areas where real-time AI processing is essential, such as unmanned and automated devices, robots, smart retail, and industrial PCs.

At CES, DEEPX will showcase its high-performance AI solutions in collaboration with global server and workstation companies such as Dell, HP, Supermicro, Lenovo, and Kaytus, as well as demonstrate innovations in robotics, smart cities, and smart with Hyundai Motor Robotics Lab, LG U+, and POSCO DX. In addition, we will unveil AI solutions applied to various SBC (Single Board Computer) boards like Raspberry Pi, Orange Pi, NXP Board, Renesas Board, and industrial PCs in collaboration with Advantec, Inventec, DFI, iEi, Biostar, etc. to high-performance, low-power, and demonstrate competitiveness in the global.

“CES 2025 is an important stage for us to showcase our achievements in mass-production verification of AI semiconductors and interoperability between various application systems with global leaders,” said Lokwon Kim, CEO of DEEPX. “Based on our ‘All-in, All-on’ strategy, we will continue to work with global partners to create a world where AI runs anytime, anywhere, and lead the global on-device AI market.”

Coverage of the DEEPX booth at CES 2025

– When: Tuesday, Jan. 7 through Friday, Jan. 10, 2025
– Location: North Hall (AI Zone), Las Vegas Convention Center, Las Vegas, USA
– DEEPX Booth Location: 9045
– DEEPX CES 2025 Teaser Video:
Dive in On-device AI

DFI and DEEPX Unveil Revolutionary Edge AI at CES 2025 to Power Smart Cities and Industries

Accelerating the real-world adoption of Edge AIoT solutions

 

DFI, a global leader in embedded motherboards and industrial computers, will join DEEPX at CES 2025 to unveil new industrial edge AI platforms integrated with the DX-M1 AI accelerator, targeting smart city applications. This collaboration reflects a shared commitment to advancing on-device and application-specific AI, paving the way for a more flexible, efficient, and sustainable edge AI adoption.

 

According to Research and Markets, the global edge AI market is projected to grow at a 20.54% CAGR, increasing from US$23.571 billion in 2024 to US$59.978 billion by 2029. To meet the growing demand for scalable and energy-efficient AI solutions, DFI integrates the DEEPX DX-M1 AI accelerator into its industrial hardware platforms, delivering cutting-edge performance tailored to modern industry needs.

 

The DEEPX DX-M1 excels with the ability to process over 16 video channels on a single chip, achieve real-time performance exceeding 30 fps with GPU-level precision, and deliver up to 25 TOPs of AI power while minimizing energy consumption. This combination of power and efficiency enables simultaneous execution of advanced AI algorithms, such as object recognition and image classification, making it a versatile solution for applications in industrial robotics, machine vision, AI-driven IPC and HPC, smart factories, and beyond.

To demonstrate this vision, DFI will showcase two fan-less embedded systems—the EC710-ASL and EC600-RPS—featuring the DEEPX DX-M1 M.2 AI accelerator. These platforms are designed to meet critical real-time demands in smart cities, enabling applications like smart transportation, surveillance, and accident prevention. The system are powered by Intel® Atom® and Intel® 14th/13th/12th Gen processors, and support Intel® Time Coordinated Computing (Intel® TCC) technology for precise real-time operations while maintaining low power consumption and minimal latency.

 

DFI continues to push the boundaries of industrial-grade embedded solutions with its latest product innovations, including the ASL253 single-board computer powered by Intel® Atom® processors, the EB100-MTU NUC system featuring Intel® Core™ Ultra processors, proprietary Out-of-Band (OOB) modules, and the versatile DFI Workload Platform. These advancements not only boost hardware performance but also enhance remote management through OOB capabilities and optimize AI efficiency to meet the growing demands of edge AIoT applications.

 

News Source – Embedded Computing Design

AI chips for all: DEEPX CEO Lokwon Kim’s vision to democratize AI technology

 

The applications of AI are becoming increasingly widespread in our daily lives. With the continuous iteration of generative AI and large language models, various innovative applications are emerging. These include, such as smart home appliances, autonomous vehicles, and robots to VR/AR applications.

 

These edge devices are equipped with AI processors that compute on the device directly, further accelerating the launch of innovative Edge AI applications. However, these devices also face several challenges, including excessive GPU power consumption and high costs limiting the widespread adoption of Edge AI products.

 

DEEPX, an AI chip startup from Korea, won three CES Innovation Awards 2024 in January for its unique AI chip ultra-gap source technology. The awards were in the categories of Computer Hardware, Embedded Technology, and Robotics, and Computer Hardware & components.

 

Their NPU processor boasts low power consumption and cost-effectiveness and addresses the issue of insufficient accuracy found in existing NPUs on the market. This technology was recognized by lots of semiconductor companies during CES 2024 and is scheduled for mass market release in late 2024.

 

 

Learn from ARM to build DEEPX into a leading On-device AI company

“The main battlegrounds in the AI era will move to the ‘edge’. Just as ARM dominated the CPU market with smartphones, the semiconductor company that dominates the edge will dominate the AI market,” DEEPX CEO Lokwon Kim said.

 

Kim, a former senior researcher in Apple’s Application Processor (AP) design, returned to South Korea to establish DEEPX after gaining extensive experience in semiconductor design at renowned companies like Broadcom, Cisco, and IBM T.J Watson. His goal is to build DEEPX into a leading AI company in the era of on-device AI, reminiscent of how ARM revolutionized the CPU market with its low-power technology.

 

On-device AI, which processes information within a mobile device without connecting to a server or cloud, is a burgeoning field. ARM broke Intel’s dominance in the CPU market with its efficient, low-power processors, which are now prevalent in smartphones and expanding into PCs and servers. Kim aspires for DEEPX to have a similar impact on the AI semiconductor industry.

 

Kim identified a weakness in Korea’s semiconductor ecosystem, particularly in system semiconductors. Drawing inspiration from Morris Chang, the founder of TSMC, who returned to Taiwan to establish a leading foundry after learning from the American semiconductor giant TI, Kim saw an opportunity to address these gaps in the South Korean market.

 

Chang predicted a demand-driven semiconductor market, leading to the creation of TSMC, which produces semiconductors on consignment. Similarly, Kim believes that the opening of the AI semiconductor market offers a chance for innovation and growth.

 

Democratizing AI Technology

Kim opposes the monopolization of AI semiconductor technology through strategic investments (SIs) that often lead to mergers, acquisitions, or alliances, resulting in technology being controlled by a single entity. His experience at Apple motivated him to advocate for more democratic and accessible AI semiconductor technology. He believes that AI semiconductors should be universally available and not restricted to proprietary use, emphasizing the importance of technology that everyone can utilize.

 

To pursue this vision, Kim established DEEPX in Pangyo, South Korea’s hub for fabless companies. He aims to create an AI semiconductor company with a unique value, promoting innovation over exclusivity.

DEEPX is positioned not as a competitor to global semiconductor giants but as a complementary force, enhancing the global semiconductor landscape by providing accessible and advanced AI technology.

 

 

Analyzing the Difference Between NPUs and GPUs: DEEPX’s Unique Solution

The necessity for developing specialized AI semiconductors, such as NPUs, stems from the limitations of GPUs, which have traditionally been used for AI computations. GPUs, originally designed to process graphic data, excel at handling large amounts of data simultaneously, making them suitable for AI learning tasks.

 

However, their high power consumption and operational costs present significant drawbacks. This makes them less ideal for “edge AI,” which involves running AI applications directly on devices like controllers, robots, and self-driving cars – collectively referred to as the “edge.”

 

NPUs (Neural Processing Units) are modeled after the human brain, offering the benefits of lower power consumption and reduced production costs. However, existing NPUs have struggled with accuracy and support for the latest AI algorithms. DEEPX stands out among fabless companies by addressing the core challenges of AI semiconductors in one comprehensive solution.

 

Unlike typical NPU vendors that release a single chip, DEEPX recognizes that different electronic devices require varying levels of semiconductor capabilities. For instance, AI for closed-circuit television (CCTV) primarily needs to analyze video, whereas AI for robots involves far more complex computations.

 

To address this, DEEPX has developed from low-end to high-end performance four chips at once: one that can connect a single electronic device for AI computations, and another that can link three or four devices for broader AI tasks. This universality across devices is a key reason DEEPX received the Innovation Award at CES 2024.

 

DEEPX’s commitment to high performance at lower power and cost also earned them the Innovation Award. The company’s latest AI algorithm, Yolo7, runs on their semiconductor DX-V1, produced using Samsung’s 28nm process.

 

This algorithm was previously incompatible with conventional NPUs. In addition, the DX-M1 chip boasts a design area one-third the size of other NPUs, and its manufacturing cost is similarly reduced by one-third. Combining low unit costs with high performance in a low-power NPU, DEEPX’s products are poised to lead the AI semiconductor market.

 

NPUs are categorized into data center-based NPUs, which handle large-scale inference, and edge-type NPUs, designed for use in electronic devices such as robots, smart cameras, smart factories, consumer electronics, etc. DEEPX targets both data center and edge NPU markets. DEEPX’s NPUs overcome the common shortcomings of existing NPUs by providing high accuracy and efficiency.

 

DEEPX’s innovation lies in creating NPUs that are not only small and cost-effective but also achieve accuracy comparable to, or even better than, GPUs. This success is attributed to DEEPX’s pioneering work in two core technologies: IQ8 (an INT8 model compression technology), and Smart Memory Access(minimizes D-RAM usage). DEEPX leads the market for low-power AI solutions, achieving the world’s highest power-to-performance ratio through proprietary advancements in hardware and software optimization.

 

Advancing Towards Mass Production

DEEPX has successfully demonstrated its original technology with sample units and is in the final stages of preparing its mass-production chip. Scheduled for market release in late 2024, the widespread adoption of products featuring DEEPX’s chips could establish the company as a technology leader in the on-device AI market by 2025.

 

The essence of on-device AI lies in low power consumption and the seamless integration of hardware and software. Since AI must operate on small devices, minimizing power consumption is crucial while maximizing AI performance within limited computing power.

 

Kim, drawing from his experience at Apple, emphasizes the importance of designing hardware and software with equal priority from the outset. Unlike Apple, which develops its own devices and services, DEEPX engaged with approximately 700 customers during product development to understand their needs and find the optimal development point. This customer-centric approach has been pivotal in refining DEEPX’s products.

 

Four companies from Israel and the U.S., including DEEPX, are vying for dominance in the emerging on-device AI market. As the market begins to flourish beyond the server market, this competition is crucial for setting future momentum.

 

Kim is confident in DEEPX’s strategy. He notes that while some competitors prioritize rapid product releases, DEEPX focuses on price and optimization first. Given the rapid evolution of AI application services, Kim believes it is more important to align with market trends and offer high-performance features at a competitive cost rather than rushing products to market.

 

DEEPX’s innovative approach to developing versatile, efficient, and cost-effective NPUs positions the company as a formidable contender in the on-device AI market. By prioritizing customer needs and maintaining a balanced focus on hardware and software optimization, DEEPX is set to lead the industry with cutting-edge AI semiconductor solutions.

 

 

DEEPX Company’s Philosophy: Valuing Technology Over Short-term Results

DEEPX is not focused on making AI semiconductors for autonomous cars or smartphones. Instead, DEEPX aims to advance the integration of AI into everyday life. The company’s products are designed to bring AI to areas such as CCTV and robots, pushing these technologies beyond outdated algorithms. DEEPX’s mission is to pioneer advancements that should have already been made in the AI semiconductor industry.

 

At the entrance of DEEPX, there is a note to employees that emphasizes the value of technology over monetary gains. It quotes Carl Sagan’s “pale blue dot,” reflecting the idea that life should be lived with value rather than chasing money and power. The message encourages the 70 employees to love their work and find it meaningful, aligning with the company’s ethos of creating technology for the greater good.

 

DEEPX believes in ‘technology that everyone shares.’ From a management perspective, AI is seen as one of humanity’s final inventions, marking the endpoint of human evolution. The goal is not to monopolize technology or chase profits but to lead the way in making AI accessible and beneficial for all. DEEPX’s AI semiconductors are the cornerstone of this vision, enabling widespread adoption of AI.

 

Future Plan: Global Expansion and Innovation

DEEPX’s visionary CEO, Kim, has outlined a strategic plan for the future:

– Global Market Entry: Starting from the second half of this year, DEEPX plans to aggressively enter the global market with its first-generation product, consisting of four AI chips. This move is set to usher in the era of “AI Everywhere.”

 

– Technological Innovation: DEEPX aims to develop new technologies that enable super-scale AI services with power consumption of less than 5W. This innovation will make advanced AI technologies more accessible and practical for widespread use.

 

– Leadership in AI: DEEPX is committed to becoming a leading comprehensive AI chip company globally. By focusing on power and cost efficiency, the company seeks to provide core technologies that transition giant AI advancements from the realm of science to everyday applications.

 

DEEPX’s mission is to integrate AI into everyday life, advancing technology in meaningful ways. By valuing innovation over profit and aiming for global leadership, DEEPX is set to play a pivotal role in the future of AI. The company’s commitment to shared technology and global expansion highlights its dedication to making AI accessible and beneficial for humanity.

 

News Source – Digitimes