Fundraising Fox

Superb AI

YC W19

San Francisco, US Β· Founded 2018 Β· Delaware corporation Β· 65 employees Β· 15 known investors

Superb AI provides an industrial AI stack for vision and analytics applications across mobility, security, defense, logistics, and manufacturing. The platform integrates data labeling, foundation models, and video analytics for real-world operating conditions.

Also known as Superb AI Inc. Β· Superb Platform

Founders & leadershipΒ· Y Combinator alumni (W19)

Superb AI was founded in 2018 by Hyun Kim, Moonsu Cha, Jonghyuk Lee, Jungkwon Lee, and Hyundong Lee.

HKHyun Kim
Hyun Kimin𝕏FounderHe previously worked as an AI engineer at SK Telecom and studied at Duke University.
MCMoonsu Cha
Moonsu ChainCo-founder and CTO
JLJonghyuk Lee
Jonghyuk LeeinCo-Founder, Software Engineer
JLJungkwon Lee
Jungkwon LeeinCo-Founder and CTO
HL
Hyundong LeeFounder

Board

JH
Jinwoong HwangBoard director

Investors Β· 15

Also in the syndicate Β· 8

DoosanDoosan InvestmentHanwha Asset Management venture fundleadHyundaiHyundai Motor GroupKakao InvestmentKT InvestmentPOSCO Investment

Funding

SEC filings, press & company announcements

$11.5M disclosed across 1 of 6 rounds Β· 2019–2025

  • Undisclosed amountPre-IPODec 2025 Β· 2 sources

    Hanwha Asset Management venture fund (lead), POSCO Investment

    Source β†—
  • Undisclosed amountSeries COct 2024 Β· 2 sources

    Doosan, Doosan Investment, Hyundai, Hyundai Motor Group, Kakao Investment, KT Investment, Premier Partners, Samsung Next

    Source β†—
β–Ά$11.5MraisedJan 2021 Β· 53 investors Β· Other
Rule 506(b)
Officers, directors & promoters on the filing
  • Jung Kwon LeeDirector
  • Jinwoong HwangDirector
  • Hyunsoo KimExecutive Officer, Director
Offering amount
$11.5M
Amount sold
$11.5M
First sale
Jan 2021
Incorporated
Corporation, Delaware, 2018
Federal exemptions
06b
Full filing on SEC EDGAR β†—

Source: SEC EDGAR Form D. Amounts as filed; amended filings shown once at their latest values.

Company profile

researched Aug 2026

Superb AI is a vision AI and MLOps company founded in 2018 that supports the full lifecycle of computer vision AI development for enterprises. Its flagship Superb Platform covers dataset creation, selection, processing, management and analysis through model training, deployment and continuous improvement, with tooling intended to be usable by staff without coding or machine learning backgrounds. The platform historically comprised modules including Superb Curate (data curation, with support for LiDAR, point cloud and sensor-fusion data), Superb Label (annotation), Superb Model (training) and Superb Services (professional services), and later added Superb Apps and Superb Video Analytics. The company works with images, video and 3D LiDAR data.

The current product stack is positioned around "industrial AI" and spans data labeling, foundation models and video analytics for mobility, security and defense, logistics, and manufacturing. In 2024-2025 Superb AI added an on-premise deployment option for customers with security or regulatory constraints, a generative AI data-generation feature for synthesizing rare or missing cases (defects, disasters, accidents), and ZERO, described as Korea's first industry-tailored vision foundation model, which is applied to varied visual tasks without extensive additional training. The company states it holds patents grouped into no-code AI (automated labeling and model training), data-centric AI (automated data curation and generation), industrial AI (industry-tailored foundation models and agents) and physical AI (3D/spatial and temporal recognition).

Founding story

Superb AI was founded in 2018 by a team including Hyun Kim, who serves as co-founder and CEO and previously worked as an AI engineer at SK Telecom from 2016 to 2018 and studied at Duke University (B.S., Ph.D. dropout). Other listed founders include Moonsu Cha, Jonghyuk Lee and Jungkwon Lee. From the outset the company focused on enterprises adopting vision AI based on images, video and 3D LiDAR, built on the premise that data quality is the essence of AI performance.

Business model

B2B software and services: enterprise licensing of the Superb Platform in cloud and on-premise configurations, plus professional services and industry-specific model and video analytics solutions. Self-serve sign-up with a free tier has also been offered alongside enterprise sales.

Enterprise software subscriptions/licensing for the platform and vision AI solutions, supplemented by professional services engagements; a free self-service entry point requires no credit card.

Traction

Superb AI reports more than 100 enterprise customers globally and over 100 global deployments, with entities in South Korea, the United States and Japan. At the Series C announcement it stated it had reached 100% of the prior full year's revenue in the first half of 2024 and was targeting annual growth above 200%. The Japan subsidiary, established in April 2023, added Toyota and Nippon Steel as customers, with the Toyota contract subsequently renewed and the Nippon Steel relationship expanding; the Japanese customer base is being extended to national research institutes and consumer electronics manufacturers. An earlier version of the site cited 5,000 machine learning teams using the product, and an 8th-anniversary post references 130 million images. Cumulative disclosed funding is stated as KRW 63 billion (about $43 million) as of December 2025. Team size is listed as 65.

Latest developments

In December 2025 Superb AI announced a KRW 14 billion (about $10 million) pre-IPO equity round joined by Hanwha Asset Management's venture fund and POSCO Investment, bringing cumulative funding to KRW 63 billion (about $43 million), with proceeds directed to AI technology development, hiring, domestic enterprise sales expansion and international growth ahead of a targeted 2026 stock market listing. The company also reports first place in the CVPR 2026 Foundational Few-Shot Object Detection Challenge with ZERO, the release of ZERO 2.2 with improved text-prompt and Korean-prompt handling, selection as a K-AI Company and participation in Korea's national AI foundation model project, designation as NVIDIA's only Physical AI partner in Korea, and a renewed Toyota contract plus expanding work with Nippon Steel.

β–ΈFull profile β€” market position, technology, go-to-market, geography, history, risks & controversies

Market position

Positions itself as an all-in-one vision AI and MLOps provider for industrial settings, citing more than 100 enterprise deployments worldwide, selection as one of Korea's "K-AI Companies" developing national AI models, participation in the Ministry of Science and ICT national AI foundation model project within an LG AI Research-led consortium as the only non-LG affiliate responsible for foundation model development, designation as NVIDIA's only Physical AI partner in Korea, and AWS Rising Star Partner of the Year. Competition awards include first place in the CVPR 2026 Foundational Few-Shot Object Detection Challenge and consecutive wins in the Hanwha Systems AI Challenge.

The company emphasizes an end-to-end pipeline in a single platform (data through deployment) rather than point tools, a data-centric approach premised on data quality determining model performance, no-code usability for non-ML staff, support for large-scale video and multi-sensor/LiDAR data, generative synthesis of scarce edge-case data, deployment flexibility including on-premise, and a patent portfolio it describes as a technical moat. ZERO is presented as reducing adoption barriers by delivering performance with relatively small data volumes.

Technology

Computer vision and MLOps technology covering automated and AI-assisted labeling, data curation without requiring separate embedding models, model training and deployment, and monitoring. Supported data types include images, video, 3D LiDAR, point cloud and multi-sensor fusion. ZERO is the company's vision foundation model for industrial environments, applied to diverse visual tasks with limited task-specific training; version ZERO 2.2 added text-prompt accuracy improvements and better Korean prompt understanding. Additional capabilities include generative AI synthesis of rare-event training data, video analytics with natural-language search and summarization over large frame volumes, and cloud or on-premise deployment. The company reports holding 38 patents across four technology groups and certifications including SOC 2 Type II and GS Level 1 software quality certification.

Go-to-market

Direct enterprise sales and demo requests, self-serve product sign-up, and channel/technology partnerships β€” the company is an AWS Differentiated Partner and was named AWS Rising Star Partner of the Year, and describes a partnership with NVIDIA covering video surveillance systems and joint delivery of AI projects at industrial sites. Several strategic corporate investors (including Hyundai Motor, Doosan and POSCO-affiliated entities) began as or became customers. Content marketing assets include a blog, community, customer stories and an on-demand academy.

Large enterprises and public institutions adopting vision AI, spanning automotive and mobility, steel and heavy manufacturing, electronics, logistics and distribution, defense, security, energy and utilities, healthcare, construction, and government research bodies. Named customers and referenced organizations include Samsung, Samsung SDS, Samsung C&T, LG Electronics, Hyundai Motor, Hyundai AutoEver, Kakao, KT, NCSOFT, Qualcomm, S-1 Corporation, Toyota, Nippon Steel, POSCO, Doosan, Hanwha Systems, Hanwha Vision, Woowa Brothers, Incheon International Airport Corporation, EVRAZ North America, Avikus, Levatas, Rowbot, Overwatch Imaging, KEPCO KPS, K-water, CS Wind, SK E&S, Korea Land and Housing Corporation, Kearney, Lemong, KRIT, the National Fire Research Institute and Ajou University Dental Hospital.

Geography

Global headquarters in Gangnam-gu, Seoul, South Korea; U.S. office in San Mateo, California; Japan office in Toranomon, Tokyo. Legal entities operate in South Korea, the United States and Japan, with customers in Korea, Japan and North America.

History

Founded in 2018, Superb AI was accepted into and graduated from Y Combinator's Winter 2019 batch and raised $2.1M in seed funding, launching Superb Platform in the same period. It raised a $9.3M Series A, was named to Forbes 30 Under 30 Asia (April 2020), became a founding member of the Global AI Infrastructure Alliance, and had a paper accepted at AAAI. A $16M Series B followed, alongside SOC 2 Type II certification and an MIT Innovators Under 35 listing. Superb AI Japan was established in April 2023, and the company was named to Forbes Korea's 50 Fast-Growing Startups 2023 and released Curate, Model and Apps modules. In 2024 it was named AWS Rising Star Partner of the Year, launched an on-premise solution, closed a $10.2M Series C (October 2024), obtained GS Level 1 software quality certification, and launched Superb Video Analytics and the ZERO vision foundation model. In August 2025 it was selected for the Ministry of Science and ICT's proprietary AI foundation model initiative, and in December 2025 it announced a roughly $10M pre-IPO round ahead of a planned 2026 listing.

Risks & controversies

The company's stated 2026 IPO plan is contingent on market conditions and not yet completed. Reported performance figures β€” including revenue growth targets, deployment counts, detection accuracy and labeling-reduction claims β€” originate from the company's own communications rather than independent verification. Revenue concentration among large Korean conglomerates, several of which are also shareholders, and reliance on partnerships with AWS and NVIDIA are further considerations.

Compiled by commissioned research from 8 cited public sources β€” announcements, filings, and press listed under research sources below.

Key figures

latest reported
Abnormal-behavior detection rate (company-stated)Jan 202692.7%
Average video search latency across 10M frames (company-stated)Jan 20262 seconds
Cumulative disclosed fundingDec 2025$43M
Defect detection accuracy (company-stated)Jan 202698%
Enterprise customersDec 2025100 customers
Global enterprise deploymentsJan 2026100 deployments
HeadcountAug 202665
Images processed (company-stated)Jan 2026130,000,000 images
Labeling reduction with ZERO (company-stated)Jan 202610 x
Machine learning teams using the platformJan 20235,000 teams
Patents (company-stated)Jan 202638 patents
Revenue growth targetOct 2024200%
Team sizeJan 202665 employees

Company-reported or press-reported figures, each dated to when it was claimed β€” not independently audited.

Competitors Β· 10

by search overlap
Roboflow106 shared keywordsRoboflow provides tools for developers to build and deploy computer vision applications, including image annotation, model benchmarking, and dataset management. The platform serves software engineers and enterprises who need to create vision-based applications.
Encord101 shared keywordsEncord provides a data platform for developing and deploying multimodal AI systems, serving AI development teams at scale across training and production environments.
Ultralytics87 shared keywordsUltralytics provides an end-to-end computer vision platform for annotating images and videos, training YOLO models on cloud infrastructure, and deploying models across 42 global regions. The platform serves developers and enterprises across industries including manufacturing, healthcare, automotive, agriculture, retail, and logistics.
Netezza77 shared keywordsIBM is a global technology company whose business spans enterprise software (including Red Hat, HashiCorp, and Confluent), IT infrastructure such as mainframes, servers, and storage, and IT consulting services. The company is also investing heavily in quantum computing and AI-based enterprise offerings, including its Lightwell open-source software security clearinghouse and the Anderon quantum wafer foundry.
Super Annotate73 shared keywordsSuperAnnotate is a data platform used to build, manage, and refine training data for AI and machine learning models. It is positioned as a data pipeline partner for teams developing AI systems.
Lightly67 shared keywordsLightly provides tools for AI teams to manage and curate training data more efficiently. The company focuses on data-centric approaches to improve machine learning model quality.
Viso65 shared keywordsViso.ai offers a computer vision platform with two products: Visonow, a tool to build vision agents from a text prompt and video, and Visosuite, an enterprise operating system to connect cameras across sites and run governed, agentic vision workflows at scale. It serves enterprises across multiple industries for use cases like monitoring, inspection, and safety alerting.
Hugging Face59 shared keywordsHugging Face is a collaboration platform that hosts and provides access to machine learning models, datasets, and applications. It offers both open-source tools for the ML community and paid compute and enterprise solutions for teams building AI applications.
Cloudfactory58 shared keywordsCloudFactory provides a platform and consulting services to develop, deploy, and operate AI systems in production, combining quality data preparation, model fine-tuning, human validation, and inference oversight. It serves enterprises in high-stakes sectors such as healthcare, finance, robotics/embodied AI, retail, logistics, manufacturing, and agriculture.
NVIDIA47 shared keywordsNVIDIA designs GPUs, AI computing hardware, and software platforms for AI development, data centers, autonomous vehicles, and robotics, serving developers, researchers, and enterprises. Its offerings include AI models, power architecture for AI factories, and compute infrastructure used across industries such as manufacturing, healthcare, and automotive.

Companies competing with Superb AI for the same Google search keywords, organic and paid, via search-intersection analysis.

Timeline Β· 22

launches, deals, and filings
Jun 2026
First place at CVPR 2026 Foundational Few-Shot Object Detection Challenge

ZERO took first place in the challenge at CVPR 2026.

source β†—

Jan 2026
ZERO 2.2 release

Update adding sharper text-prompt accuracy and improved Korean prompt understanding.

source β†—

Dec 2025
Superb AI raises about $10M pre-IPO round ahead of planned 2026 IPO

KRW 14 billion (approximately $10 million) pre-IPO equity round joined by Hanwha Asset Management's venture fund and POSCO Investment; cumulative funding stated as KRW 63 billion (about $43 million).

$10M source β†—

Aug 2025
Selected as elite team for Korea's Proprietary AI Foundation Model initiative

Selected by the Ministry of Science and ICT; participates in the national AI foundation model project within an LG AI Research-led consortium as the only non-LG affiliate responsible for foundation model development.

source β†—

Jun 2025
CVPR 2025 challenge placements

2nd place in the Object Instance Detection Challenge and 4th place in the Foundational Few-Shot Object Detection Challenge at CVPR 2025.

source β†—

May 2025
1st Place, Hanwha Systems IR Instance Segmentation Challenge 2025

source β†—

Jan 2025
NVIDIA partnership on video surveillance systems

Technical collaboration with NVIDIA on video surveillance systems (VSS) and joint delivery of AI projects at global industrial sites; company describes itself as NVIDIA's only Physical AI partner in Korea.

source β†—

Jan 2025
Toyota contract renewal

Superb AI renewed its contract with Toyota and continues expanding its partnership with Nippon Steel, its first Japanese customer.

source β†—

Jan 2025
Launch of ZERO vision foundation model and Superb Video Analytics

ZERO is described as Korea's first industry-tailored/industrial vision foundation model for real-world environments; Superb Video Analytics launched in the same period.

source β†—

Nov 2024
AWS Rising Star Partner of the Year

source β†—

Nov 2024
IITP commendation for contributions to ICT technology dissemination

source β†—

Oct 2024
Superb AI closes $10.2M Series C

Series C round with new investment from Doosan Investment, Hyundai Motor Group, Kakao Investment and Samsung Next, plus existing investors KT Investment (fourth participation) and Premier Partners (third participation), bringing cumulative investment to $36.9 million.

$10.2M source β†—

Sep 2024
1st Place, Hanwha Systems IR Object Detection Challenge 2024

source β†—

Jan 2024
Launch of Superb On-Premise Solution

On-premise deployment option for clients with strict security requirements or regulatory constraints preventing use of public networks or cloud services.

source β†—

Jan 2024
Launch of generative AI-based data generation feature

Feature enabling creation of ground-truth quality data for rare instances such as defective products, natural disasters or accidents.

source β†—

Jan 2024
Acquired GS Level 1 software quality certification

source β†—

Apr 2023
Establishment of Superb AI Japan

Japanese subsidiary established, subsequently securing customers including Toyota and Nippon Steel.

source β†—

Dec 2022
Seoul Metropolitan Government Mayor's Commendation in Technology Development

source β†—

Jan 2021
Acquired SOC 2 Type II certification

source β†—

Apr 2020
Named to Forbes 30 Under 30 Asia

source β†—

Jan 2019
Graduated from Y Combinator Winter 2019 batch

source β†—

Jan 2019
Launch of Superb Platform

source β†—

Dated company events from announcements, filings, and press; legal rows summarize public dockets and regulator releases.

Legal entities Β· 1

corporate structure
Superb AIDelaware

In the news

β–ΈResearch sources Β· 8

primary sources listed

8 public sources were cited for this profile; the first-party ones are listed here.

Frequently asked questions

What does Superb AI do?
Korean-founded vision AI company offering an end-to-end MLOps platform, a vision foundation model and video analytics for industry.
Who founded Superb AI?
Superb AI was founded by Hyun Kim, Moonsu Cha, Jonghyuk Lee, Jungkwon Lee, Hyundong Lee in 2018.
Who are Superb AI's investors?
Superb AI's investors include Atinum Investment Co. Ltd., Duke Capital Partners, KT investment Inc., Premier Partners, Samsung Next, Y Combinator, Pegasus Tech Ventures.
How much funding has Superb AI raised?
Superb AI has disclosed $11.5M raised across 1 of its 6 known rounds.
Where is Superb AI headquartered?
Superb AI is headquartered in San Francisco, US.