INFORMATION ASYMMETRY · AI, PHYSICAL AI & EDUCATION

How Chinese Universities Are Adapting to AI, AI+ and Physical AI

A source-based guide to embodied-intelligence research and entrepreneurship at leading universities, major redesign at ordinary institutions, and practice-led change in vocational education.

Official ministry, government and university sources · Updated August 1, 2026
현대 중국 대학 캠퍼스와 로봇 연구실, 직업교육 실습장에서 학생들이 피지컬 AI 하드웨어를 제작하는 수묵화
Start with the terminology

Physical AI is used here as a broad explanatory term. Chinese policies and universities more often use embodied intelligence, intelligent robots, humanoid robots, intelligent manufacturing and unmanned systems. They overlap but are not exact synonyms.

Read regional innovation networks, not universities alone

Beijing, the Yangtze River Delta and the Greater Bay Area provide different research, manufacturing and startup environments. Locations do not imply a ranking.

Beijing · UCAS / TsinghuaShanghai / Zhejiang · ShanghaiTech / Westlake / ZhejiangGuangdong · SCUT / SUSTech
Shared province-level map projected from one coordinate system at one scale · Taiwan included · Color indicates relevant province-level units and may not match the detailed boundaries of a city cluster or event location.Unified province-level boundary data · china-map-geojson / ISC license · Albers equal-area conic projection · Mainland China, Hainan, Hong Kong, Macau and Taiwan are rendered in one projection rather than composited from separate maps.

Six concepts to keep separate

具身智能

Embodied intelligence

Integrates perception, reasoning and action through interaction between a body and its environment; overlaps with, but is not identical to, Physical AI.

人工智能+ / AI+X

AI+ / AI+X

Embeds AI into existing fields such as engineering, materials, medicine, transport, management and the arts.

微专业

Micro-program

Usually a short, non-degree cluster of courses or skills credential; credit and assessment vary by institution.

现代产业学院

Modern industry college

A university-industry platform for joint curriculum, mentors, facilities and projects—not usually an independent university.

人形机器人

Humanoid robot

A hardware form factor resembling the human body; shape alone does not prove embodied intelligence or autonomy.

未来机器人

Future robotics

An interdisciplinary major in China’s undergraduate catalogue; delivery capability must still be checked institution by institution.

The shift is broader than adding AI majors

AI literacy, major redesign, micro-programs, interdisciplinary institutes, hardware projects and commercialization are moving together at different levels.

01

AI literacy for all

Common learning in AI principles, limits, ethics, data and disciplinary use.

02

Major-portfolio redesign

Uses new, resized, suspended and abolished programs to fit mission and regional demand.

03

AI+X and micro-programs

Adds AI, robotics and data skills without requiring students to change degrees.

04

Interdisciplinary institutes

Connects mechanics, control, electronics, computing, materials and cognition around embodied systems.

05

Projects and hardware practice

Trains sensing, control, embedded systems, vision, safety and integration on real hardware.

06

Transfer and faculty startups

Links lab work, pilot testing, validation, transfer and startups while preserving legal distinctions.

Seven universities with visible research and startup ecosystems

We distinguish prototypes, faculty startups, university-project incubation, research-institute spinouts and corporate joint labs.

01

南方科技大学

Southern University of Science and Technology

Shenzhen

A visible chain from AI literacy and formal majors to robotics research and a faculty startup.

02

华南理工大学

South China University of Technology

Guangzhou

Its strength is engineering-wide AI+ integration and industry-project education.

03

西湖大学

Westlake University

Hangzhou

Combines small-cohort early research, AI+Science and lab-level embodied intelligence.

04

中国科学院大学

University of Chinese Academy of Sciences

Beijing and the CAS institute network

Its core model combines degree education with the Chinese Academy of Sciences institute network.

05

上海科技大学

ShanghaiTech University

Shanghai

Links an AI honors track, the STAR robotics center and a broader transfer system.

06

清华大学

Tsinghua University

Beijing

Runs core AI research, university-wide AI+ education and a new embodied-intelligence institute in parallel.

07

浙江大学

Zhejiang University

Hangzhou

Combines AI+X curricula with joint engineer training on real robots.

How ordinary four-year universities are pivoting

The pattern is portfolio rebalancing based on mission, regional industry and employment—not a blanket conversion to AI degrees.

01

Add and close programs together

Rebalances programs around regional industry, demand, employment, faculty and facilities.

Provincial universities proposed AI, intelligent-manufacturing and low-altitude programs alongside closures; proposals, approvals and enrollment are distinct.
02

Common AI plus disciplinary AI

Provides common literacy first, then adapts data, tools and ethics to each discipline.

Zhejiang Normal University offers a university-wide AI introduction; national policy calls for common plus specialized AI courses.
03

Use micro-programs for rapid reskilling

Adds interdisciplinary skills faster than redesigning a full degree.

The MOE’s Double-Thousand program reported 2,654 micro-programs and about 74,000 2025 graduates; these are generally non-degree supplements.
04

Rebuild existing majors with AI+

Changes data, labs, projects and assessment inside existing engineering and non-engineering degrees.

Changsha University of Science and Technology grouped four majors around industrial embodied intelligence; Shanghai Institute of Technology pursues both AI+ and +AI.

Major-portfolio adjustment in public data

New, suspended and abolished programs move together, and not all adjustment is caused by AI.

2024年度1,839

신규 전공 설치점

2024年度2,220

모집정지 전공 설치점

2024年度1,428

폐지 전공 설치점

2025目录845

일반대 학부 전공

2026目录883

일반대 학부 전공

2026批准9

구신지능 전공 첫 승인 대학

How vocational institutions are preparing

Their focus is closer to installation, commissioning, integration, operation, maintenance and deployment, supported by teachers, labs, industry projects and safety systems.

01

中等职业教育

Secondary vocational education

Builds fundamentals in mechanics, electronics, safety, operation and basic maintenance.Connects automated equipment, sensors, PLCs and basic robot operation to work procedures
02

高职专科

Higher vocational college

Emphasizes installation, commissioning, integration, operations, maintenance and deployment.Intelligent robot technology, industrial robot technology and applied AI
03

职业本科

Vocational bachelor

Targets systems integration, process improvement, applied innovation, technology transfer and field engineering.Intelligent manufacturing engineering, robotics and newly added embodied-intelligence engineering

A five-level test for the phrase ‘university startup’

A university name in a company profile does not prove equity ownership or direct incubation. Contracts, registration and official records matter.

01

Corporate joint research

A joint lab, course or equipment relationship does not mean the university created the firm.

ShanghaiTech × Fourier Intelligence
02

Faculty or alumni startup

The founder-university affiliation is verified; equity or direct incubation needs separate evidence.

Tsinghua–Spirit AI; Zhejiang–DeepRobotics
03

University-project incubation

The university officially states that the company grew from a university project or research.

SUSTech project–LimX Dynamics
04

Affiliated-institute spinout

An institute in the same system incubated the company, but the university legal entity may be different.

CAS Shenyang Institute–SIASUN
05

University equity or direct subsidiary

The strongest claim, requiring registry, disclosure or investment documents proving ownership or control.

Do not use this label without verified evidence

Twelve questions for evaluating new programs and ecosystems

Curriculum, faculty, equipment, real projects and graduate outcomes matter more than program names and MOU counts.

  1. 01신고·제안, 교육부 승인, 실제 모집, 재학생 교육, 첫 졸업을 구분했는가
  2. 02정규 학위전공과 마이크로전공·단기과정·실험반을 구분했는가
  3. 03기계·전자·제어·임베디드·센서·비전·안전·통합 교과가 함께 있는가
  4. 04전임교원과 기업 겸임멘토의 역할·시간·평가책임이 명확한가
  5. 05로봇·센서·연산장비의 수량보다 학생 접근시간과 가동률이 공개되는가
  6. 06시뮬레이션을 넘어 실제 하드웨어 조립·고장·안전·유지보수를 훈련하는가
  7. 07캡스톤이 전시용 시제품인지 실제 기업 문제·검수·유상과제로 이어지는가
  8. 08논문·특허·협약 수와 기술이전·계약·납입·매출을 분리했는가
  9. 09교수 창업, 대학 부화, 연구소 스핀오프, 공동연구기업을 구분했는가
  10. 10졸업률·전공일치 취업·고용유지·기업평가·창업생존을 추적하는가
  11. 11신설 전공이 과잉공급될 때 정원조정·폐지와 재학생 보호절차가 있는가
  12. 12한국에서 적용할 때 국내 산업수요·안전·표준·법·윤리를 다시 검증했는가

KOREA RESPONSE

How Korea should respond

Translate verified facts into actions for public bodies, industry, universities, research, finance, end users and professional services.

중앙정부·지방정부·관련기관인재정책·지역 교육인프라
Korea response direction (proposed)

전교생 AI 문해력–비학위 마이크로자격–교원·장비·수요가 검증된 정규 피지컬 AI 학위의 3층 구조를 설계합니다. 대학마다 로봇을 중복 구매하기보다 권역별 공동 실증실·안전시험장·데이터 수집장·디지털트윈과 실물 장비를 결합한 개방형 기반을 조성합니다. 신설·과잉·개편 필요 전공 목록과 재학생 보호계획을 주기적으로 공개합니다.

Evidence and indicators

확인자료 · 지역산업 인력수요, 전공별 정원·교원·장비·취업, 공동시설 운영규정 / 지표 · 전공일치 취업, 공동장비 가동률, 참여기업·유상과제, 교육비용 대비 성과

대기업·중견·중소·스타트업산학과제·채용·장비생태계
Korea response direction (proposed)

일회성 장비기증보다 실제 공정문제, 데이터, 현장멘토, 검수기준과 채용경로를 공동교과에 제공합니다. 대기업은 폐쇄형 벤더 교육이 되지 않도록 중소 장비·부품·소프트웨어 기업이 참여할 인터페이스를 열고, 중소·스타트업은 모듈 단위 캡스톤과 유상 인턴·도제를 통해 수요를 검증합니다.

Evidence and indicators

확인자료 · 문제정의서, 공동교과·IP·데이터 계약, 멘토 투입시간, 채용·구매 절차 / 지표 · 캡스톤 현장적용, 유급인턴, 채용·구매 전환, 공동개발 매출

대학·전문대·직업대·산학협력단교육과정·교원·학생보호
Korea response direction (proposed)

연구중심대는 모델·알고리즘·아키텍처, 응용대는 통합·검증·기업과제, 전문대는 설치·운영·정비·현장배치, 직업학부형 과정은 시스템통합·공정개선·기술이전으로 역할을 나눕니다. 모든 과정은 기계·전자·제어·임베디드·센서·비전·로봇안전·시스템공학을 AI 위에 함께 쌓고, 교원 기업연수와 기업엔지니어 공동지도를 제도화합니다.

Evidence and indicators

확인자료 · 교과맵, 전임·겸임교원 역량, 학생 장비접근시간, 안전·IP·연구윤리 규정 / 지표 · 역량평가, 프로젝트 재현성, 전공일치 취업·고용유지, 중도탈락·안전사고

출연연·전문연구기관공통기술·시험평가·표준
Korea response direction (proposed)

대학이 반복 구축하기 어려운 로봇 안전, 성능측정, 데이터 품질, 센서·액추에이터·제어, 인간–로봇 상호작용의 공통 시험법과 참조과제를 제공합니다. 중국 대학의 논문·시제품은 동일 조건의 재현시험으로 평가하고, 국내 산업현장에 맞는 공개 벤치마크와 안전 프로토콜을 축적합니다.

Evidence and indicators

확인자료 · 시험절차·데이터셋·장비교정·윤리심의·표준 초안 / 지표 · 재현시험 통과, 시험기간·비용, 표준·인증 반영, 기업 공동실증

정책금융·은행·VC·PE교육인프라·대학기술 투자
Korea response direction (proposed)

학과·연구원 간판이나 교수 경력만으로 투자하지 않고 대학 프로젝트 부화, 교수 창업, 연구소 스핀오프, 외부기업 공동연구, 대학 지분기업을 분리해 실사합니다. 시설금융은 장비구매보다 공동활용률·기업과제·교육수입·유지보수 재원을 보고, 창업투자는 IP 소유·실시권·교원겸직·학생권리와 후속 고객을 확인합니다.

Evidence and indicators

확인자료 · 등기·지분·기술이전·겸직·IP 계약, 장비 운영수지, 고객검증 / 지표 · 후속투자·매출, 장비수익·가동률, IP분쟁, 기술이전 회수

실수요기업·조달기관현장실증·구매·인재평가
Korea response direction (proposed)

로봇 데모나 대회 수상보다 작업 성공률, 안전정지, 고장복구, 환경변화 적응, 사람 개입시간과 총소유비용을 공개합니다. 학생·연구팀의 과제를 실제 현장에 넣되 무상 PoC를 반복하지 않고, 합격기준을 충족하면 유상실증–구매–유지보수 계약으로 이어지는 관문을 만듭니다.

Evidence and indicators

확인자료 · 현장시험계획·로그·안전·SLA·구매기준 / 지표 · 임무 성공률, 가동·복구시간, 사람 개입, 유상전환·재구매

교육개발·사업화기관·법률·회계·특허교과·창업·국제협력 실행
Korea response direction (proposed)

중국어 전공명과 기관 성격을 한국어로 정확히 번역하고 정규학위·마이크로전공·현대산업학원·공동연구실을 구분합니다. 한중 공동교육·연구·창업에서는 학생·교수·대학·기업의 발명자와 권리자, 데이터 반출, 논문공개, 특허출원, 장비책임과 분쟁절차를 프로젝트 시작 전에 문서화합니다.

Evidence and indicators

확인자료 · 중문 원문·학위 인가·공동연구·기술이전·데이터·IP 계약 / 지표 · 번역·권리 불일치 예방, 계약기간, 기술이전·창업 생존, 분쟁건수

Primary official sources

01교육부 · 2024년도 학부전공 조정 결과와 2025 전공목록02교육부 · 고등교육 전공 조정 최적화 정책03교육부 · 교육 디지털화와 ‘공통+특색’ AI교과04교육부 · 취업역량 향상 쌍천계획05교육부 · 2026 인공지능+교육 행동계획06교육부 · 2026 일반대 학부전공 목록07베이항대 · 구신지능 전공 첫 승인 대학 9곳08남방과기대 · AI 교육과 전공 신설09남방과기대 · LimX Dynamics 산학 프로젝트 부화10남방과기대 · AI고등연구원11남방과기대 · 로봇연구원12화난이공대 · 혁신반·AI+ 전공 교육13화난이공대 · 인간요인지능·장비시스템 공동실험실14서호대학 · AI 학부전공15서호대학 · AI+과학 교과16서호대학 · 기계지능실험실 구신지능17중국과학원대학 · AI학원 참여 연구소·지능로봇18중국과학원대학 · 첫 AI 학부 졸업생19상하이과기대 · AI 영예과정20상하이과기대 · STAR 로봇센터21칭화대 · AI+ 학부 공통교육22칭화대 · 구신지능·로봇연구원23저장대 · AI STEP 교육체계24저장대 · DeepRobotics 공동양성기지25교육부 · 2021 직업교육 전공목록26교육부 · 758개 직업교육 전공 교육표준27교육부 · 2026 직업교육 전공목록 증보28교육부 · 현장엔지니어 특별양성29교육부 · 쌍사형 교원 인정기준30헤이룽장성 · 구신지능 로봇 직업대회·교과 재구성
Back to Information Asymmetry