SenseTime Spins Off AI Healthcare Company, Raises 1 Billion in Six Months to Target the "Medical World Model
SenseTime Medical: From Spin-off to Near-Unicorn in Just Six Months
SenseTime has spun off an AI-powered medical company that has quickly risen to near-unicorn status in half a year.
Recently, the core executive team made their first public appearance, presenting a comprehensive roadmap and business layout.
---
Vision and Strategic Goals
SenseTime Medical strives to be:
- Designer and enabler of "future hospitals" powered by medical foundation models
- Builder of a world model for the medical domain, capable of complete perception and deep understanding
- Creator of a self-learning, continuously evolving smart hospital ecosystem
---
Rapid Fundraising & Strategic Backing
- Early investment from Midea Group’s YinFeng Holdings and Renwei Technology (People’s Medical Publishing House Group)
- Additional hundreds of millions RMB round with Lenovo Capital, Legend Capital, NineThemes Capital, Shenran Investment
- Series A launched — total capital raised in 6 months: 1 billion RMB
Xu Li, Chairman & CEO of SenseTime Group:
> SenseTime Medical is a core extension of the group’s “1+X” strategic ecosystem.
> The group will remain its most solid technical backing and steadfast strategic partner.
---
Technology Architecture
Approach: “General-specialized integration”
On top, a central brain orchestrates multi-modal foundation models — pathology, imaging, endoscopy — coupled with a knowledge base and scenario toolsets.
Core model: Self-developed medical large language model DaYi®
Performance:
- Benchmarked across 8 dimensions (13,191 test items) — ranked #1 overall
- Outperforms DeepSeek (full), GPT-5, and other general models
---
Technical Highlights
- High-quality Medical Corpus
- 250B+ tokens (~400B Chinese characters), 100+ medical disciplines
- Continuous tracking of new high-value data and efficacy optimization
- Industrial-grade RAG Framework
- Custom-built for medical scenarios
- Source verification and knowledge checks before answer generation
- Clinical Reasoning Enhancements
- Complex data extraction from patient records/lab reports
- Long-text comprehension for multi-year histories
- “Long-thought” datasets combining expert and evidence-based knowledge
---
Dual Middleware Platforms
1. DaYi® Bots Intelligent Agent Development Platform
- No-code orchestration, private knowledge base conversion, unified model scheduling
- Standardized products: DaYi ZhiJian, DaYi ZhiYan, DaYi YiZhu
2. Multi-modal Foundation Model Application Production Platform
- For medical imaging: algorithm library + deep learning framework
- Full-lifecycle, no-code data annotation → training → evaluation
Result: “Data–Model–Scenario–Industry” closed loop
---
SenseCare® Smart Hospital Solution
Includes:
- Smart Clinical Diagnosis Platform
- Smart Patient Services Platform
- Smart Medical Research Platform
- Smart Medical Cloud Platform
Over 40 AI modules deployed in 10+ specialties:
- Lung
- Coronary arteries
- Head & neck blood vessels
- Liver
- Musculoskeletal systems
---
Pathology Success Story
SenseCare® Intelligent Pathology AI-Assisted Diagnosis System:
- 30–50% work efficiency improvement
- Nearly 100% tumor detection sensitivity
- 80% benign case negative predictive rate
---
Surgical Planning Breakthroughs
SenseCare® Intelligent Surgical Planning System:
- Automatic 3D reconstruction, lesion detection, multi-option surgical planning
- Real-time interactive simulation & analysis
- IDEAL — China’s first liver surgery AI multimodal planner, launched with Academician Dong Jiahong’s team
---
Patient Services: "DaYi ZhiJian" & Mini-App
DaYi ZhiJian:
- Triage guidance, pre-consultation, AR navigation, cloud accompaniment, follow-up
SenseTime DaYi Mini-App:
- Medication consultation
- Report interpretation
- Lab analysis
- Personal health record management
---
Research Assistant: "DaYi ZhiYan"
Capabilities:
- Topic selection
- Literature interpretation & translation
- Q&A, sample size estimation
- Paper writing
- Statistical chart generation
- Manuscript review
- Polishing
Purpose: Free doctors from tedious tasks and boost research productivity
---
Executive Team
CEO: Zhang Shaoting
CMO: Qian Kun
COO: Shen Tian
CFO: Zhao Du
CSO: Shang Hailong
VP of R&D: Wang Rui
---
Talent Composition
- ~100 members; 70% in product & R&D
- Alumni from Stanford, Harvard, CUHK, Tsinghua, Peking, SJTU
- Blend of medical + engineering expertise
---
CEO Insights: Medical Software Evolution
Zhang Shaoting’s 3 stages:
- Project stage — digitize manual processes
- Product replication — deep learning for high-frequency problems
- Model capability (3.0) — vertical specialization in medical knowledge
Skipping stage 2 limits large models to shallow hospital applications.
---
Commercial Achievements
Ruijin Hospital:
- Liver Surgery Intelligent Planning → 3D reconstruction in 2–5 minutes
- Assisted 400+ complex liver resections
---
Roche Collaboration
Xiao Luo Zhi Duo Xing platform:
- Covers 90 cities, 700 tertiary hospitals
- Supports 4,000+ doctors, 600+ research projects
- Saves 20,000+ collective hours
---
National-Scale AI Medical Data Facility
Shanghai Shenkang Hospital Development Center:
- Largest clinical-application-oriented big data infrastructure in China
- Unified compute + algorithm + data for low-code/no-code model building
---
Expanded Deployments
- Integrated with Lenovo SSG product channels
- You Hospital (Midea Group) — 15 AI diagnostic apps launched
---
AI Healthcare & Monetization
Platforms like AiToEarn官网:
- Open-source, global, cross-platform content generation & monetization
- Publish simultaneously to Douyin, Kwai, WeChat, Bilibili, Rednote, YouTube, etc.
- See: AiToEarn博客 | AiToEarn文档 | AI模型排名
---
International Expansion
- Singapore: First AI+healthcare device license; hospital rollout
- Indonesia: First business deal secured
---
Future: Building the "World Model"
Goal: Dynamic simulation of medical environments, not just Q&A.
Capabilities:
- Anatomical + physiological simulation
- Predict treatment risks & effects
- Act as a “digital laboratory” for hospitals
---
SenseTime's “1+X” Strategy
Core “1” = Generative AI + Visual AI
“X” = Innovation units — intelligent driving, smart healthcare, robotics, retail
---
Spatial Intelligence Research
Collaboration with Nanyang Technological University:
- Puffin Model integrates camera geometry into multi-modal AI
- Enables optimized camera understanding & generation
Zhang Shaoting & Chen Change Loy → Program Chairs at CVPR 2026
---
Closing Note
AI in healthcare and spatial intelligence continues to expand.
Platforms like AiToEarn官网 will help bridge cutting-edge AI research with global adoption and monetization.
---
Would you like me to also create a condensed one-page executive summary of these highlights so stakeholders can grasp the milestones at a glance?
