Z Venture Capital Invests in Config Intelligence to Drive Robot Foundation Models

Z Venture Capital (ZVC) has announced its investment in Config, a platform company addressing the data bottleneck in robotics. Config utilizes its unique 'Action Estimator' technology to transform human videos into high-precision motion data, powering its hardware-agnostic foundation model 'CFG-1'.
資金調達NQ 93/100出典:PR Times

📋 Article Processing Timeline

  • 📰 Published: May 22, 2026 at 21:04
  • 🔍 Collected: May 22, 2026 at 12:31
  • 🤖 AI Analyzed: May 27, 2026 at 07:08 (114h 37m after Collected)
Z Venture Capital (ZVC) has invested in Config, a startup specializing in the development of Robot Foundation Models.

### Investment Background: The 'GPT-2.5 Moment' for Robotics
The field of robotics is currently facing a transformation similar to the period just before ChatGPT appeared for LLMs. While hardware and architectures are maturing, the critical missing piece is 'data.' Unlike text or images that can be scraped from the web, high-quality robotic motion data is scarce, hindering the application of scaling laws to physical AI.

### Config's Breakthrough Solution
Config is an end-to-end platform company that handles everything from data collection to model building. Its proprietary 'Action Estimator' technology converts low-cost human activity videos into high-precision robotic motion data with sub-millimeter accuracy (approx. 500 micrometers). This effectively solves the industry's largest bottleneck: data availability.

### Key Achievements and Strengths
- **Proprietary Model 'CFG-1'**: Trained on over 100,000 hours of unique data, this model is hardware-agnostic and already generating revenue.
- **Strategic Partnerships with Korean Giants**: Collaborating with Samsung, LG, and Hyundai, Config leverages South Korea's world-leading robot density and manufacturing base to gather unique data that competitors cannot easily replicate.
- **Elite Founding Team**: Led by CEO Minjoon Seo (ex-Meta FAIR) and CTO Hyungmok Son (ex-Waymo), the team combines expertise from top global AI and autonomous driving institutions.

### Future Outlook
The market for robot foundation models and physical AI data platforms is projected to grow at over 50% annually, exceeding $18 billion (approx. 2.7 trillion JPY) by 2030. Config is positioned to secure a dominant first-mover advantage by building an insurmountable moat of proprietary data.

FAQ

Config はどのような技術を持つ企業ですか?

ロボット基盤モデルの開発に必要なデータ収集からモデル構築までを一貫して提供するプラットフォーム企業です。特に、人の作業映像を高精度なロボット動作データに変換する「Action Estimator」技術に強みがあります。

ロボット基盤モデル「CFG-1」の特徴は何ですか?

10万時間以上の独自データで学習されており、特定のハードウェアに依存することなく動作する汎用性を持っています。既に実際の収益化も実現しています。

Config が韓国市場で優位に立っている理由は何ですか?

Samsung、LG、Hyundai といった韓国の大手企業と提携しており、世界屈指の製造業基盤を持つ韓国の工場環境から、他社が模倣困難な独自の動作データを蓄積できるためです。

ロボット基盤モデルの市場展望はどうなっていますか?

フィジカルAI領域で最も急成長している分野の一つであり、2030年までに年率50%超の成長を続け、市場規模は180億ドル(約2.7兆円)を超えると予測されています。

Config の創業チームにはどのような人物がいますか?

CEOのMinjoon Seo氏(Meta FAIR出身)、CTOのHyungmok Son氏(Waymo出身)、Chief ScientistのKimin Lee氏(Google出身)など、AI・自動運転のトップランナーが名を連ねています。