Return of Industrial-Software PhDs: CAD/CAE/EDA, Industrial AI and OS Are Becoming a Hard Gap
(Zhengxianling®Full-Path Guide for High-Level Returnees / Vol. 16 / Timeline format)
Abstract
This guide (Vol. 16, Timeline format) of the Zhengxianling®five-step returnee-decision series addresses overseas PhDs, postdocs and senior R&D engineers in industrial software and foundational software (CAD/CAE/EDA), computational mathematics and numerical simulation, IC design methodology, computer/software engineering and control science. It rests on three converging structural signals - autonomous-controllability becoming the key-research mainline, export controls turning usable into must-use, and AI restructuring design tools - and argues that the real opening lies in building the tools at the kernel level (geometry kernel, PDE solvers, EDA algorithms, real-time schedulers), not in using the tools; the window period is the bargaining period. The article translates the five-step method (positioning, city, platform, application, offer) into a 12-month calendar: four path-fit profiles (academic, enterprise R&D, entrepreneurship, postdoctoral), national and provincial talent-program matrices (coded QM / HJ / CC / RJ; QM, backed by enterprises, carries the highest weight here), compensation and city-platform logic, and an offer negotiation broken into seven items. All program figures, age lines and application windows are cross-checked against the 2026 official notices but roll over annually and must be re-verified against current guides; all monetary amounts are subject to official current announcements.
Keywords: industrial software; CAD; CAE; EDA; industrial AI; industrial OS; numerical simulation; talent programs; Zhengxianling®
Opening: if you start preparing now, what to do each month in the next 12 months
This article does not give you a job list but a calendar. Industrial software is a track where the closer to the kernel, the scarcer: many people can use software, but very few can write out, line by line, a geometry kernel, a partial-differential-equation solver, a layout-verification algorithm or a real-time scheduling kernel. Precisely because of this scarcity, returning home is not a linear submit-resume-wait-for-offer action but a timeline that must be scheduled 12 months ahead.
This article uses the timeline / action-oriented framework: from an action-coach perspective it unpacks the five-step method (positioning, city, platform, application, offer) in time order - position 12 months before returning; pick city and lock platforms 6 months before; prepare application and negotiate offer 3 to 6 months before; restart advanced applications after landing. You just follow the calendar forward.
But action presupposes a window: is the industrial-software window real? Yes, and the window itself is the bargaining period - below are three hard signals first, then we return to the timeline.
I. Step 0: judging the window - three structural signals industrial software is undergoing
The first step is neither submitting a resume nor picking a city, but confirming whether this track is worth returning to now. Industrial software offers three stacked signals whose intersection is exactly the overseas PhD core competency area.
(I) First signal: autonomous-controllability becomes the key-research mainline; R&D-design tools are the weakest link
Domestic industrial software shows marked tiered localization: per CCID Consulting 2025 China Industrial Software Development Report, overall localization is about 35%, with management-type (ERP, etc.) about 70%, production-control (MES/DCS/PLC) about 52%-60% and R&D-design (CAD/CAE/EDA) about 10%-26% (large cross-agency spread); at the sub-level CAD about 20%, CAE below 10%, EDA below 10% (the MIIT Fifth Electronics Industry Institute 2025 report likewise puts EDA localization below 10%); the high-end market remains foreign-led and R&D-design tools are the hardest bottleneck. These ratios are consultancy/research-institution estimates with multiple coexisting calibers, must follow the current edition of each source report, and are not used as a commitment. MIIT Guidelines on Equipment Renewal and Technological Upgrading in Key Industries (工信厅规〔2024〕33号, 20 Sep 2024) sets a target of renewing about 2 million units of industrial software and 800,000 units of industrial OS by 2027, focusing on OS/databases/middleware, on CAD/CAE/CAM/MES/ERP, and on PLC/DCS/SCADA/SIS/embedded software, covering 27 key industries (figures and lists per the MIIT original). For a PhD, this means bottom-layer kernel roles are persistently pulled by policy for years.
(II) Second signal: export controls turn usable into must-use, pulling kernel talent back
External controls turn domestic substitution of R&D-design tools from a choice into a compulsory question: access and upgrades of advanced-node EDA, high-end multiphysics simulation and complex-surface 3D kernels are restricted; per industry research reports, the U.S. has in recent years extended key-software export controls into industrial software (specific items and entity lists per the current announcement of the U.S. Department of Commerce BIS; this article makes no precise-date claim), pushing downstream customers from passive trial to batch procurement and primary supply, with leading domestic vendors raising both capacity and R&D. The tighter the control, the greater the need for people who have actually built solvers, chained EDA flows end-to-end or written geometry-kernel operators - exactly what overseas PhDs accumulate in labs and vendor pre-research.
(III) Third signal: AI is restructuring design tools, opening a new industrial-AI job group
On 28 Apr 2026 the General Office of MIIT and the Comprehensive Department of the National Data Administration jointly issued the Notice on the Joint 2026 Model-Data Resonance Action (工信厅联科函〔2026〕193号), driving AI-model and data-resource co-promotion across 20 key industries (steel, petrochemical, machine tools, autos, medical equipment, power equipment, ships, aerospace, electronic components, software, telecom, cybersecurity, etc.), targeting a benign data-model-application loop by end-2026. Surrogate models accelerating simulation, AI-assisted place-and-route, generative 3D modeling and digital-twin coupling make composite PhDs who know both physical mechanisms/engineering constraints and data/algorithms structurally scarce. In system-level modeling/simulation, fully self-owned domestic platforms already serve major engineering programs - substitution is leaping from usable to easy-to-use.
(IV) Supplementary hard signals: named 2026 policy calibers
The 15th Five-Year Plan (十五五, 2026-2030) outlines raising the level of foundational and industrial software and building internationally competitive digital industry clusters, and fully implementing an AI-plus action; MIIT AI+Software Special Action Implementation Plan (工信部信发〔2026〕209号, issued 2 Sep 2026, press conference 11 Sep) sets AI-assisted and AI-native fusion paths, deep-merging CAD/CAE/EDA with AI, deploying coverage of 20,000 large software enterprises and 100 intelligent-agent benchmarks by 2028 and 15+ adaptation centers in foundational software. The statement that foundational software is the cornerstone of the digital society and industrial software is the soul of manufacturing digital transformation was made by an MIIT information-development official at the State Council Information Office Good Start of the 15th FYP press conference (26 Aug 2026) - a different session from the 11 Sep launch, listed separately here. Traceable baseline: per MIIT 2025 Software Industry Operations (30 Jan 2026), 2025 software revenue was 15,483.1 bn CNY (+13.2%), of which software-product revenue 3,236.1 bn (+10.4%), industrial-software product revenue 333.0 bn (+9.7%), foundational-software revenue 214.6 bn (+11.1%) and IC-design revenue 442.1 bn (+18.9%) - the last directly relevant to EDA demand. Per IDC China Core Industrial Software and Industrial AI Forecast 2026-2030, Chinas core industrial-software market (CAD/CAE/EDA/PLM/MES/APS) grows from 40.88 bn CNY (2025) to 84.31 bn (2030), CAGR about 15.6%, with the AI-plus-industrial-software sub-market compounding at about 49.0% from 5.30 bn to 38.94 bn and rising from 13.0% to 46.2% share. R&D-design tools are a low share of industrial software (industry estimate about 8.5%), the hardest to localize and the scarcest in kernel-level PhDs.
Three factors sum to this article core judgment: the opportunity for returning industrial-software PhDs is not in using tools but in building tools at the kernel; the window is the bargaining period. Product-side incentives and 0-to-1 evidence:
(1) first-version software incentives continue, with localities granting roughly 20% of actual sales for first-version software, up to 3.0 mn CNY per enterprise, and about 20% of R&D spend for industrial apps (Inner Mongolia 2024 policy list item 52 applies this; Hubei grants up to 10.0 mn CNY bilateral reward per its Industrial Software Action Plan 2024-2026);
(2) a domestic university team self-built geometry kernel went from v1.0 (2021) to v5.0 (2025), deployed in aircraft/autos/ships/machine tools, proving bottom-layer kernel substitution is engineering-feasible; (3) per the Blue Book on Domestic Industrial Simulation Software 2025-2026, domestic simulation lags foreign on 10-million-mesh parallel solving by 30%-60%, leading firms annual R&D is under 200 mn CNY, and stable profitability typically takes 8-10 years - the gap is both a shortfall and the source of a PhD hard-kernel premium, but entrepreneurship or early joining must budget for long cycles and slow payment.
(V) Core-fact verification table (source / traceability grade / item by item)
Traceability grades: publicly traceable = official document / authoritative release / listed-company announcement, with number and caliber findable publicly; per public report = engineering progress quoted from authoritative media, some dates not independently verified; industry caliber, reference only = salary ranges etc. without a single official caliber, synthesized from public job info and industry reports - not to be treated as a commitment or precise figure.
Caliber distinction (added in revision): three mutually non-convertible statistics exist - (1) MIIT industrial-software product revenue (enterprise-report caliber, 2025 333.0 bn CNY); (2) consultancy industrial-software market size (incl. integration/services, wider scope, e.g. CCID); (3) IDC core industrial software (limited to CAD/CAE/EDA/PLM/MES/APS, 2025 40.88 bn). Their definitions differ; never convert or compare directly; each citation marks the publisher and caliber. Baseline discipline: industry data here is a fact-layer supplement beyond the shared content-baseline (SSOT) authorization; before cross-article use the platform must write it back to the baseline as a series-shared caliber to avoid cross-article inconsistency.
# | Key fact/data | Source and caliber | Traceability | Status |
1 | By 2027, large-scale renewal of industrial software and industrial OS (CAD/CAE/CAM/MES/ERP/PLC/DCS/SCADA/SIS and embedded OS as focus) | MIIT Guidelines 工信厅规〔2024〕33号 | Publicly traceable | Verified (figures per original) |
2 | R&D-design localization about 10%-26% (CCID about 26%, others about 10%-15%; sub CAD about 20%, CAE<10%, EDA<10%); overall about 35%, management about 70%, production-control about 52%-60% | CCID 2025 report; MIIT 5th Electronics Institute (EDA<10%); multiple bodies coexist | Industry caliber (estimates), multi-caliber, reference only | Direction consistent; the earlier R&D-design under-10%-overall phrasing was corrected to a range |
3 | EDA localization below 10% (MIIT 5th Institute); other bodies put overall 15%-18%, analog>30%, digital-implementation<8%, advanced-node<5% | MIIT 5th Institute 2025; Industrial Software Alliance etc. | Industry caliber, multi-caliber | Direction consistent; cite agency and year |
4 | CAE localization rose from about 7% (2016) to about 16.2% (2022, Zhiyan); CCID 2025 puts CAE below 10% - different scope/year, do not compare directly | Zhiyan (2022); CCID (2025) | Per public report; caliber diff noted | Direction consistent; conflict noted |
5 | A leading domestic 3D-CAD vendor ranks among the top domestic brands by market share; 2D/3D kernel self-coded | Enterprise disclosure / industry (entity categorized) | Per public report | Direction consistent; no precise ranking |
6 | Three coexisting market-size calibers: (1) MIIT product revenue 333.0 bn (2025, +9.7%); (2) consultancy size about 364.97 bn (2024); (3) IDC core industrial software 40.88 bn (2025) | MIIT Operations (2026-01-30); CCID; IDC | Publicly traceable (MIIT) + estimates | Verified (MIIT); three calibers not comparable |
7 | Key-software new talent gap by 2025 about 830k, of which industrial software about 120k (forecast); another expert public forecast about 200k over five years, demand growth over 50%/yr | Key-Software Talent Demand Forecast (2021); public talks | Per report; forecast | Direction consistent; the earlier composite-gap-over-500k (no source) was deleted |
8 | Only about 30 universities offer industrial-software majors; MOE added 29 majors incl. industrial software; CAE-algorithm and EDA-architecture roles nationwide below 5,000 | Public research and media (2026) | Industry caliber, reference only | Direction consistent; earlier 164k workforce (no unified official stat) replaced |
9 | EDA high-end composite talent gap is large, training long, master-plus composite share low (values are agency-survey) | Public research | Industry caliber, reference only | Direction consistent; no precise value |
10 | System-level modeling/simulation already shows a fully self-owned domestic platform used in major programs | Enterprise/media (platform categorized) | Per public report | Direction consistent |
11 | 28 Apr 2026 MIIT and NDA issued the 2026 Model-Data Resonance Notice (工信厅联科函〔2026〕193号) across 20 industries, loop target by end-2026 | MIIT/NDA (2026-04-28) | Publicly traceable | Verified (doc no. and list confirmed) |
12 | Advanced-node full-flow EDA in export controls; customers shift from trial to batch procurement | Industry media | Per public report | Direction consistent |
13 | Big Fund III tilts toward EDA bottleneck; many localities set EDA procurement/R&D subsidies | Industry media (amounts per current announcement) | Industry caliber, reference only | Direction consistent |
14 | In the 15th FYP many provinces strengthen industrial software: Hubei global industrial-software capital; Jiangsu/Liaoning focus on CAE/EDA | Provincial 15th-FYP public texts (2026) | Publicly traceable | Verified |
15 | Overseas Outstanding Youth 1.0-3.0 mn CNY/3yr; born on/after 1 Jan 1986; 36+ months continuous overseas teaching/research post after the PhD; not yet returned, or returned on/after 1 Jan 2025; after funding must resign overseas job and work full-time in China for 3 years | NSFC Overseas Outstanding Youth 2026 Guide (Mar 2026) | Publicly traceable | Verified; birth-year line rolls annually, do not reuse |
16 | Boxin: 280k CNY/yr (national 2-yr term) + one-off 80k postdoc fund; born on/after 1 Jan 1994, PhD within 3 years; returnee and foreign PhDs ineligible; in-service postdocs may not apply; about 500 funded in 2026 | National Postdoc Committee national-funded program (Tier A) 2026 notice and multiple host-unit forwards (consistent) | Publicly traceable | Verified; birth-year line rolls annually |
17 | Returnee high-level study-abroad talent subsidy: one-off 300k; Chinese nationality, <=50, PhD abroad, returned after 2021 and now full-time; those already in other national programs excluded; many localities match 1:1 | MOHRSS current notice and local HR bureau forwards (e.g. Xiamen) | Publicly traceable | Verified |
18 | China Postdoc Science Fund general 80k (natural science, 2026 plan about 300), special 180k (2026 plan about 1,200); in-service not eligible | Postdoc Science Fund Guide 2026 (中博基字〔2026〕1号) | Publicly traceable | Verified |
19 | 2025 software revenue 15,483.1 bn (+13.2%), industrial software 333.0 bn (+9.7%), foundational 214.6 bn (+11.1%) | MIIT | Publicly traceable | Verified |
20 | AI+Software Plan (工信部信发〔2026〕209号, issued 2 Sep 2026, press conference 11 Sep) sets AI-assisted and AI-native paths; 20k large SW firms, 100 agent benchmarks, 15+ adaptation centers by 2028 | MIIT (信发〔2026〕209号) and press-conf record | Publicly traceable | Verified (doc no. and year confirmed) |
21 | Foundational software = cornerstone, industrial software = soul (MIIT info-dev official); 2025 industry revenue 15.48 tn CNY, 6.2x 2012, CAGR 15.1% | SCIO Good Start of the 15th FYP (2026-08-26); MIIT AI+Software launch (2026-09-11) | Publicly traceable | Verified; the soul statement was originally mis-tagged to the Sep 11 launch; now listed by the two sessions |
22 | Core industrial software 40.88 bn (2025) to 84.31 bn (2030), CAGR about 15.6%; AI-plus-industrial-software 5.30 to 38.94 bn, CAGR about 49.0%, share 13.0% to 46.2% | IDC China Core Industrial Software and Industrial AI Forecast 2026-2030 (Aug 2026) | Per public report; agency forecast | Revised to IDC current caliber; the earlier 2024-2029 CAGR about 41.4% / 9% to 22% did not match and was removed |
23 | 2025 software structure (MIIT): embedded-system SW 1,386.9 bn (+9.3%), product 3,236.1 bn (+10.4%, incl. industrial 333.0 bn, foundational 214.6 bn), IT services 10,636.6 bn (+14.7%); four-class split puts R&D-design share low (industry estimate about 8.5%) | MIIT Operations (2026-01-30); four-class share is an industry estimate | Publicly traceable (MIIT) + industry caliber | Verified; sub-share must be tagged as an industry estimate |
24 | EDA market: three international giants over 70% combined; local leaders about 5%-6% | IDC etc. competitive landscape (entities categorized) | Per public report; agency | Direction consistent; no precise share |
25 | Domestic simulation lags foreign 30%-60% on nonlinear/10M-mesh solving; leading firms R&D under 200 mn/yr; typical profitability takes 8-10 yr | Blue Book on Domestic Industrial Simulation Software 2025-2026 | Per public report; agency | Direction consistent; compilation-unit estimate, not official |
26 | An industrial-software park opened 22 Oct 2025, total investment 2.15 bn CNY, 260k sqm, 5.0 bn CNY fund, about 50 firms to move in (about 70% industrial software), target over 100 firms and at least 2.5 bn revenue by 2027 | Park-opening public report (local caliber) | Per public report | Direction consistent |
27 | Geometry-kernel 0-to-1: a university team self-built kernel v1.0 (2021) to v5.0 (2025), deployed in aircraft/autos/ships/machine tools (team categorized) | University and media | Per public report | Direction consistent |
28 | First-version incentives (local examples): Inner Mongolia subsidizes 20% of actual sales, up to 3.0 mn per enterprise, and 20% of industrial-app R&D up to 3.0 mn; Hubei gives up to 10.0 mn bilateral reward | Inner Mongolia 2024 policy list item 52; Hubei Action Plan 2024-2026 | Publicly traceable (provincial) | Verified (per local current document) |
29 | Annualized key-research lists: MIIT and seven other departments AI+Manufacturing Opinion (Jan 2026) lists industrial-software intelligentization; Hunan implementation rules (Apr 2026, self R&D at least 5.0 mn, 2 yr max); Hubei up to 10.0 mn per core project; Chongqing 15th FYP strengthens first-version certification and professional-software matrix | Ministries and provincial public docs | Publicly traceable | Verified (Hunan/Hubei per public doc) |
30 | Application windows (2026): Overseas Outstanding Youth closed 15 Jun 2026; Boxin online 24 Feb to 24 Mar; Postdoc Fund 79th general and 19th special 1 to 31 Mar; returnee subsidy local recommendation by early March - all roll annually | NSFC, Postdoc Committee, Fund Council, MOHRSS and local current notices | Publicly traceable | Verified |
31 | Hubei 15th FYP proposes building the global industrial-software capital; per the provincial econ-info dept, 2025 Hubei industrial-software core scale 35.4 bn CNY, +12.9%, first in central China | Hubei 15th FYP text; provincial dept data (via media) | Traceable (plan) / per report (scale) | Verified; the global-capital statement located to provincial-plan level |
32 | Liaoning 15th FYP states raising autonomous controllability of industrial CAE and industrial-control software; provincial key special led by Dalian University of Technology develops an aerospace solver validated on major programs | Liaoning 15th FYP text; Liaoning Dept of Industry and IT proposal reply (2026-06-25) | Publicly traceable | Verified |
33 | Jiangsu 15th FYP seeks decisive breakthroughs in foundational/industrial software; issued the Jiangsu AI-plus Industrial-Software Breakthrough Action Plan 2026-2028 | Jiangsu 15th FYP text; provincial industry-IT public doc | Publicly traceable | Verified |
34 | Claim that in Oct 2025 the U.S. extended key-software export controls to industrial software | Industry research-report citation | To verify; no precise citation | Per report; items and list per current BIS announcement |
One-line conclusion: the policy document numbers (工信厅规〔2024〕33号, 工信厅联科函〔2026〕193号, 工信部信发〔2026〕209号), MIIT operation data and provincial 15th-FYP statements are publicly traceable; talent-program age lines, funding standards and windows were checked against the 2026 current notices but all roll annually and must be re-verified before release/declaration. Localization rates, market sizes and talent gaps are agency estimates/forecasts with multiple calibers, all tagged reference-only, never used as hard commitments.
II. Your 12-month overview: turning the five-step method into a calendar
The overview below maps the five-step method onto a timeline. Remember one principle: every step must precede the job-hunting action - positioning and city judgement before applying, host-employer qualification before onboarding, or the window closes before you do.
Phase | Five-step | Core monthly action | Key output |
T-12 months | 1 Positioning | Four-path self-test; list kernel/solver/EDA/industrial-AI results and project-delivery records | One main + one backup path; verifiable results list |
T-6 to 9 months | 2 City + 3 Platform | Lock 2-3 cities by industry-fit>platform-density>host>hukou; list 10-20 target units | Graded target-unit list + host-qualification pre-judgement |
T-3 to 6 months | 4 Application (start) | Confirm host for QM/Overseas Youth/returnee subsidy; start CSC-certification materials | Host intent + application-window timetable |
T-3 months | 5 Offer | Seven-item negotiation; distinguish R&D vs engineering track; check service/restitution clauses | Written offer (terms in contract) |
Landing to +3 months | 4 Application (land) | Complete onboarding/CSC certification; start provincial programs; confirm national contact | First provincial wave + national contact scheduled |
One reminder: placing host-employer confirmation at T-6 months is the most irreversible step of the timeline - returning first and then finding a unit lets windows such as Overseas Outstanding Youth close before you.
III. T-12 months positioning: four paths and the industrial-software PhD capability profile
Path divergence for industrial-software PhDs is more kernel-driven than for AI: the one judgment is whether you build tools or use them, whether you do frontier algorithm/kernel or engineering delivery. Hence four paths: academic, enterprise R&D (this article core), entrepreneurship and postdoctoral.
(I) Four-role fit matrix
Path | Suitable people | Core opportunity and host | Key action | Fitting programs |
Academic (university/institute) | Control theory, computational math, numerics, EDA algorithms; strong papers but limited engineering delivery | Strong math/mechanics/CS universities, computational-math/engineering-mechanics institutes, national numerical-simulation platforms, new R&D bodies | Ask whether there are industry/program tasks and student engineering practice (pure academia detached from scenarios limits later steps) | Overseas Outstanding Youth, provincial youth programs, study-abroad return research-start fund |
Enterprise R&D (build tools) core here | Has built geometry kernel/solver/EDA algorithm/real-time kernel, with industrial delivery record | Domestic CAD/CAE/EDA vendors, national numerical-simulation center, industrial OS/real-time DB vendors, industrial-AI platforms | Contact talent office; confirm R&D track, toolchain/compute, whether host for QM/CC | QM, CC, local industry high-level talent |
Entrepreneurship | Productizable kernel/solver/EDA point-tool/industrial-AI module | Industrial-software parks, overseas-returnee parks, concept-validation and pilot platforms | First complete at least one real industrial POC | Study-abroad entrepreneurship start program, Chunhui Cup, HICOOL, startup guarantee loans |
Postdoctoral (buffer) | Need a buffer, accumulating domestic industrial scenarios and application conditions | University mobile stations (academic), enterprise workstations (industry, key) | Choose a workstation with localization key-research tasks | Boxin (returnee PhDs ineligible), Postdoc Science Fund, overseas-talent postdoc special |
One conclusion: the optimal main path for most industrial-software PhDs is enterprise R&D - building tools; the academic path suits algorithm/numerics, and must pick a platform tightly bound to program/industry tasks.
(II) Capability-label profile (read as a capability purchase order; direct match first)
Direction | PhD profile - capability labels (direct experience) | Preferred background | Paper/result preference |
3D-CAD geometry kernel | Solid/surface modeling kernel (B-Rep), boolean and tolerance algorithms, constraint solving, complex-surface accuracy and robustness, C++ low-level, data-format compatibility | Computational geometry/graphics + mechanical engineering + kernel development | Kernel operators and patents > papers |
CAE multiphysics solver | FEM/BEM/meshless/isogeometric, structural/fluid/EM/thermal coupling, large-scale parallel, benchmarking against mainstream commercial solvers, Fortran/C++, V&V | Computational math/mechanics/aerospace/naval + solver development | Algorithm + engineering cases > pure theory |
EDA design & verification | Full-flow chaining (synthesis/place-route/timing signoff/physical verification), algorithm and PDK fit, RTL simulation | IC design methodology/CS algorithms + process understanding | Full-flow ability + point-tool breakthrough > single paper |
Industrial OS/real-time DB | Real-time scheduling kernel, microkernel/virtualization, industrial protocol stack, high-reliability redundancy | OS/embedded/control science | Kernel code/patents > papers |
Industrial AI x simulation | Surrogate/reduced-order models, AI-assisted layout, generative modeling, digital twin | ML + physics-simulation composite | Delivery case + algorithm > pure paper |
One conclusion: institutes/vendors buy your hands-on experience at a specific technology node, not the most papers - when positioning, align yourself to the one or two hardest labels.
(III) Positioning self-check
Self-check question | Direction |
Is the core result paper-type or convertible/kernel-type? | Academic vs enterprise/entrepreneurship |
Have you built geometry kernel/solver/EDA algorithm/real-time kernel? | Enterprise build-tools path feasibility |
Continuous overseas research 36+ months after the PhD? | Feasibility of Overseas-Youth academic channel |
How many years to the target program age line? | Academic/youth window |
How much income volatility and failure risk can you bear? | Entrepreneurship feasibility |
Do you need the security of establishment/tenure? | Academic vs enterprise/entrepreneurship |
Any hard family constraint on city (education/health/spouse)? | Region choice |
Does your sub-direction have a mature domestic host (kernel vendor/institute/EDA)? | Industry-path feasibility |
(IV) Four-path differentiated risks (irreversible decision points)
Path | Main risk | Irreversible point |
Academic | Few pure-theory posts; later applications/promotion limited if detached from industry | Overseas Youth requires not-yet-fully-returned - onboarding first may already forfeit eligibility |
Enterprise R&D | Easily recruited as application/ops rather than kernel R&D; platform delivery varies | Not confirming R&D track and host qualification before joining forfeits bargaining power |
Entrepreneurship | Long customer decisions, slow payment, on-site validation; heavy R&D | No POC/prototype before returning means high probability of no orders |
Postdoctoral | Mistaking a buffer for a destination; exit-retention and direction not bound | Not confirming whether the workstation has industry tasks forfeits the gain of engineering experience |
(V) Path-combination strategy (advanced)
Combination | Note |
Enterprise institute + QM/CC | Best for engineering/industrialization: firm provides job and salary, QM/CC provide policy backing and host |
University post + Overseas Youth | Best for numerics/EDA algorithms; highest academic value, locks tenure and matching |
Postdoc + youth program | Enter an enterprise workstation to keep flexibility while applying to the Postdoc Fund and provincial youth programs |
Senior flexible hiring | Not full-time: joint lab, offshore innovation, talent enclave to take localization tasks |
Key reminder (a general rule; full text in Chapter VI): talent-program application must be hosted by an employer/platform and is never paid automatically.
IV. Talent gap and salary signals: which roles compete for PhDs (why it is worth coming)
The scarcity here is not too few people but wrong capability: universities train by single discipline, CS lacks industrial scenario, mechanics/mechanical lack algorithms, those who know mechanism do not code and those who code do not know mechanism. The most urgent need is people who bridge both and can go down to the kernel layer.
(I) Sample in-recruiting PhD/senior roles (read as a capability purchase order; entities categorized, subject to current job ads)
Role | Capability | Note (subject to current) |
National numerical-simulation center - solid-mechanics CAE R&D | Mechanics/aerospace/mechanical/computational-math PhD; FEM/X-FEM/meshless/isogeometric; static/heat/vibration/impact/fatigue/optimize; C/C++, Fortran, Linux | PhD-led, open recruitment, interview; major-equipment engineering application |
Leading domestic 3D-CAD vendor - geometry kernel R&D | Computational geometry/graphics; B-Rep kernel, boolean, constraint solving, complex-surface robustness; C++ low-level | Kernel role, needs portfolio/code proof |
EDA vendor - full-flow/point-tool algorithm | IC design methodology/CS algorithms; synthesis/place-route/timing/physical verification; PDK fit | Full-flow chaining experience carries highest premium |
Industrial OS/real-time DB vendor - real-time kernel | OS/embedded/control; real-time scheduling, microkernel/virtualization, industrial protocol stack | High-reliability redundancy; security note |
Industrial-AI platform - simulation surrogate/industrial algorithm | ML + physics-simulation composite; surrogate/reduced-order, AI-assisted design | Mechanism x data composite carries highest premium |
(II) Three-tier salary reference (ranges are industry caliber, reference only, not a commitment)
Tier | Profile | Range (10k CNY/yr, reference) | Note |
Base | Fresh overseas PhD entering kernel/solver/EDA algorithm role | about 30-45 | Landing point for most industry R&D roles; per offer |
Premium | Kernel/solver/EDA full-flow real-combat experience | about 45-70+ | Leading vendors and defense/energy carry notable premium |
Fresh-grad discount | No delivery/kernel result for fresh PhD | below base | Starting point clearly below premium; anchor on lowest tier + verifiable matching |
Expectation management: the above are reference values for PhDs with industrial-kernel/systems experience; fresh graduates without delivery results start clearly lower. Ranges synthesize public job info and industry salary reports; per actual offer and official salary announcement - this guide is not a salary commitment.
Caliber note (added in revision): the above is a role range for PhDs with kernel/solver/EDA real-combat experience, differing from whole-industry statistics and not directly comparable. Per industry public statistics, the 2024 average monthly salary for industrial-software talent was about 11.6k CNY, most annual salaries in 150k-300k CNY (incl. non-PhD and engineering track), while internet-software roles averaged about 20.3k CNY monthly. Together they show the premium comes from kernel capability + R&D track rather than the industry average; if an offer sits only at the industry average, return to the three questions - kernel role, R&D track, host-application - and renegotiate.
V. T-6 months city and platform: by the cluster logic of building tools
The city logic differs slightly from AI: it weighs kernel/simulation R&D-platform density plus downstream manufacturing scenarios, not subsidies alone. Selection order follows the baseline: industry-fit > platform-density > host > hukou. All amounts below are per official current announcement.
Platform-anchoring supplement: besides vendors and institutes, industrial software has two high-value hosts - (1) industrial-software parks, e.g. a park opened Oct 2025 (investment 2.15 bn CNY, 260k sqm, 5.0 bn fund, about 50 firms, about 70% industrial software) offering rent/compute/fund in one for PhD startups; (2) a national research body foundational-software center (public establishment, PhD priority, recruiting numerical-algorithm and architecture roles). When choosing, list local industrial-software parks or national foundational-software key-research platforms as a plus; note that key-research lists are annualized (open competitions, provincial key-software specials, first-version certification are organized annually) so windows roll and are per the current notice of the competent authority.
(I) Four-tier city model (industrial-software view)
Tier | Cities | Core feature | Suitable talent |
1 | Beijing, Shanghai, Shenzhen, Guangzhou | National numerical-simulation/EDA/CAD institutes and top-vendor HQs, dense central-SOE institutes, strongest research and capital networks | Top scholars for national programs, kernel/EDA senior leaders, hardtech startups |
2 | Nanjing, Suzhou, Hangzhou, Hefei, Wuxi, Wuhan, Qingdao, Jinan, Xiamen | Industrial-software/EDA-strong cities, dense scenarios, clear delivery, strong overseas parks and pilot platforms | R&D backbone, Overseas-Youth candidates, growing founders, postdocs |
3 | Chengdu, Chongqing, Xian, Changsha, Zhengzhou, Tianjin | Low cost, active platforms; many provinces target CAE/EDA/embedded in the 15th FYP | University posts, regional industry talent, work-life balance |
4 | Dalian, Harbin, Changchun, Shenyang, Fuzou, Nannang, Kunmang, Guiyang | Local specialty industries or policy lowlands, low competition | Specialty-discipline talent, low-cost startup and life |
Tier-4 city industrial-software notes (per local current policy and job ads):
Dalian: deep embedded-software and Japan-outsourcing base, hosts industrial OS/embedded roles; low cost.
Harbin: university CS/control accumulation, hosts industrial-software algorithm and embedded directions; fits academic/postdoc transition.
Changchun: automotive-software and embedded-control scenarios, demand for industrial OS/automotive software.
Shenyang: equipment-manufacturing intelligentization, CAE/simulation landing space in equipment (security note).
Fuzou/Nannang/Kunmang/Guiyang: local specialty industries and policy lowlands, few kernel roles, better as low-cost options; check each role and host qualification.
(II) Industrial-software specialty cities (by track)
City | Industrial/foundational-software specialty | Fitting direction |
Beijing | National numerical-simulation platform, central-SOE institutes, dense EDA/industrial-AI HQs | CAE solver, EDA, industrial OS, industrial AI |
Shanghai | EDA full flow and IC design methodology, foreign R&D centers plus local innovation, Lin-gang/Zhangjiang smart-manufacturing specials | EDA, CAE, digital twin, industrial AI |
Shenzhen | EDA point-tool breakthroughs, hardtech startup ecosystem, local EDA procurement/R&D specials | EDA, industrial AI, entrepreneurship |
Wuhan | 15th FYP global industrial-software capital, universities plus Optics Valley, photonics/industrial-software fusion | CAE, industrial software, industrial AI |
Nanjing/Suzhou | Jiangsu 15th FYP CAE/EDA hard breakthrough, IC and industrial-software strong cities, many hosts | CAE, EDA, industrial software |
Hefei | Quantum/IC/university-of-science platforms, EDA and algorithm talent with university backing | EDA, industrial AI, algorithms |
Wuxi | IC/IoT/high-end equipment, dense industrial-software and EDA scenarios | EDA, industrial software, embedded |
Chengdu/Chongqing | Sichuan-Chongqing embedded/automotive software; Chongqing strengthens first-version certification | Embedded, industrial OS, automotive software |
Xian | Hardtech and defense-automation background, aerospace simulation demand | CAE, industrial OS (security/political-vetting note) |
Changsha | Construction-machinery intelligentization plus domestic CAE/simulation demand | CAE, simulation, digital twin |
Tianjin | 15th FYP foundational/industrial-software R&D and integration, Xinchuang and scenario resources | Industry application software, management software, embedded |
Shenyang/Dalian | Liaoning 15th FYP raises industrial CAE and control software; provincial special led by Dalian Univ. of Tech with aerospace solver validated on programs | CAE solver, industrial-control software (security note) |
Jinan/Qingdao | Shandong 15th FYP breaks through high-end and embedded industrial software, builds national key-software park and China Software Famous City | High-end industrial software, embedded |
Xiamen/Fuzou | Fujian 2026 software-tech-innovation grants at 300/200/100 ten-thousand tiers (per current doc) | Industrial software, domestic-substitution research |
Harbin/Changchun | University CS/control accumulation; automotive software and embedded-control scenarios (security/vetting per unit rule) | Algorithms, embedded, automotive software, academic and postdoc |
(III) Target cities by direction
R&D direction | First-apply region | Note |
3D-CAD kernel | Beijing, Shanghai, Wuhan, Nanjing | Kernel-platform firms plus strong-university joint R&D |
CAE multiphysics solver | Beijing, a national numerical-simulation center in southwest, Dalian, Changsha, Nanjing | Institutes plus heavy-industry/aerospace simulation demand |
EDA full-flow/point-tool | Shanghai, Shenzhen, Beijing, Nanjing, Hefei | Advanced-node controlled, so domestic chaining demand is most rigid |
Industrial OS/real-time DB | Beijing, Xian, Chengdu, Shenzhen | Energy/industrial-control scenarios, security note |
Industrial AI x simulation | Shanghai, Shenzhen, Hangzhou, Hefei, Wuhan | Dense digital-economy and compute platforms |
(IV) 30-second city quick card
What you do | Look first | Why |
Kernel/solver/EDA algorithm | Platform density + whether it hosts QM/CC | Determines both building tools and applying to industry programs |
Industrial AI x simulation | Compute/data platform + manufacturing scenarios | Mechanism-x-data composite delivery |
Entrepreneurship | Overseas park + local EDA/industrial-software special + first-version procurement | Guidance and subsidies decide payment and validation |
Academic | Control/math/CS strong universities + program tasks | Academic path limited if detached from industry scenarios |
(V) Four-dimension qualitative city scoring (de-pseudo-precision note)
The table uses strong/mid/weak to characterize, a synthesis of public talent policy and industry caliber, not an official ranking; all hukou/subsidy/application calibers are per the current notice of local talent offices and competent authorities.
City | Industry scenario | R&D platform | Host | Hukou friendly | Fitting direction |
Beijing | Strong | Strong | Mid | Weak | CAE/EDA/industrial OS/industrial AI |
Shanghai | Mid | Mid | Mid | Strong | EDA/CAE/digital twin |
Shenzhen | Mid | Mid | Weak | Strong | EDA/industrial AI/startup |
Wuhan | Mid | Weak | Mid | Mid | CAE/industrial software (global capital) |
Nanjing | Mid | Mid | Mid | Mid | CAE/EDA/industrial software |
Suzhou | Mid | Mid | Mid | Mid | IC/industrial software/EDA |
Hefei | Weak | Mid | Mid | Mid | EDA/industrial AI/algorithms |
Wuxi | Mid | Mid | Weak | Mid | EDA/industrial software/embedded |
Hangzhou | Weak | Weak | Mid | Mid | Industrial AI/digital twin |
Chengdu | Weak | Mid | Weak | Strong | Embedded/industrial OS/automotive software |
Xian | Weak | Mid | Weak | Strong | CAE/industrial OS (security note) |
Changsha | Mid | Mid | Weak | Strong | CAE/construction-machinery intelligentization/digital twin |
(VI) Differentiated host checks + three checks + four job channels
Host type | Industrial-software specific focus | Fitting path |
Academic | Math/mechanics/CS discipline rating; program/industry tasks; student engineering practice; Overseas-Youth application support | Academic, postdoc |
Industry (vendor/institute) | R&D vs engineering track; whether kernel/solver/EDA product line is self-owned; toolchain/compute/real data; whether hosts QM/CC | Industry (core here) |
Postdoc | Whether the workstation has a localization key-research task; exit-retention rate; supervisor engineering resources; special application success rate | Postdoc |
Startup | Industrial-software park/overseas park; rent/compute/fund; first-version/Xinchuang warehousing guidance; real POC resources | Entrepreneurship |
Three-check principle: one, check historical delivery (were terms honored in contract, application success rate); two, check match depth (is core business closely aligned to your kernel/solver/EDA direction, with real data/program tasks/compute); three, check application willingness (will the unit invest resources to assist, with a dedicated liaison).
Channel | Path |
Official/semi-official | MOE CSC, Western Returned Scholars Association Homecoming Plan, local HR/talent bureaus, overseas parks |
University/institute direct | Target HR office, school site, numerical-simulation/EDA groups, institute recruitment pages (PhD roles often skip job platforms) |
Enterprise direct and fairs | Domestic industrial-software/EDA vendors, central-SOE institutes sites; Zhengxianling® and similar overseas-talent fairs |
Supervisor/alumni/academic network | Overseas supervisors, domestic collaborators, alumni, conferences, mobile-station contacts (often better than open ads); Zhengxianling® aggregates scattered supervisor/alumni/collaborator relations into directed invitations and role matching |
One conclusion: finding a qualified host (especially an industry host for QM/CC) directly decides application success, mattering more than a city subsidy amount.
(VII) Non-cluster provincial-capital quick check (family-placement spillover scenario)
Industrial-software PhD jobs cluster in kernel/simulation cities, but hard family constraints (spouse job, children education, health) often push candidates to non-cluster provincial capitals. The table gives the fit logic: kernel/solver/EDA roles are generally scarce here, mainly hosted via provincial universities, Xinchuang firms, digitalization service providers, compute infrastructure and government digital projects - fitting family-first candidates or no-hat roles. All amounts and establishment calibers are per the local current talent and recruitment announcements.
City | Fit logic (kernels scarce, focus on application and host) | Host and fitting direction |
Shijiazhuang/Hebei | Xinchuang and electronic-info industry, digital-industry research; few kernels | Provincial universities, Xinchuang firms, digitalization service providers; industry application software, embedded |
Taiyuan/Shanxi | Energy industrial-internet and Xinchuang scenarios | Provincial universities control/CS, energy digitalization providers; industry application software, embedded |
Hohhot/Inner Mongolia | Big-data and compute base, government digital integration | Compute infrastructure, government integrators; industry application software |
Lanzhou/Gansu | Northwest software and equipment-manufacturing simulation demand | Lanzhou University and other institutes; CAE application, embedded |
Xining/Qinghai | Renewable-energy digitalization, green-power scenarios | Energy digitalization providers; industry application software |
Yinchuan/Ningxia | National integrated compute-network hub node | Compute and data infrastructure; industrial AI, compute |
Urumqi/Xinjiang | Frontier informatization and energy digitalization | Regional energy/IT service providers; industry application software, embedded |
Haikou/Hainan | Free-trade-port digital economy, talent-attraction policy | Digital-economy parks, attraction policy; industry application software |
Nannang/Guangxi | Digital-economy cooperation toward ASEAN | China-ASEAN information port, digitalization providers; industry application software, embedded |
Guiyang/Guizhou | Big-data industry highland | Compute and data service providers, high-tech zone; industrial AI, compute |
Lhasa/Tibet | Aid-frontier special, establishment and title green channel | Mainly application and O&M; industry application software |
VI. T-3 to 6 months application: the valuable channels for industrial-software PhDs
Application is the core lever for bargaining power. The peculiarity here is that enterprise channels (QM, CC) weigh more than pure academic ones (Overseas Youth), because most build-tools opportunities are on the enterprise R&D side and institute engineering platforms.
(I) General rules (the single full-text source)
# | Rule | Meaning |
1 | Application must be hosted by an employer/platform | Overseas Youth, QM, CC, returnee subsidy, postdoc types all need a university/institute/firm/park to host; never auto-paid |
2 | Academic first-get-a-slot-then-apply does not work | Overseas Youth requires not-yet-fully-returned; joining first may forfeit; never return then slowly find a unit |
3 | All amounts/conditions per official current announcement | Hukou/subsidy/funding/salary change fast; all figures here are public calibers, final per competent authority, employer agreement and current notice |
(II) National talent-program matrix (industrial-software fit order)
Program (code) | Host | Fit note | Key caliber (per current) |
QM | Enterprise/institute | Best for industrialization, engineering tech, kernel/platform development | Led by MIIT, enterprise-hosted, youth/innovation/entrepreneurship categories; need not fully return during application; national + local matching |
Overseas Outstanding Youth | University/institute | Only for numerical-method/EDA-algorithm academic paths | 1.0-3.0 mn CNY/3yr; born on/after 1 Jan 1986; 36+ months continuous overseas teaching/research post after PhD; not yet returned or returned on/after 1 Jan 2025; after funding resign overseas job and work full-time in China 3+ years. Age line rolls annually per NSFC current guide |
CC | Central/state SOE | Major-engineering research; domestic CAE/EDA kernel substitution is a typical fit | National + SOE matching, per current notice |
HJ | Enterprise host | An industry project, not a personal title; for tech-transfer | Project-based funding, per current notice |
RJ | Banking/financial system | A finance-system talent program; industrial-software PhDs mostly not applicable, may skip | Special funding, per current notice |
Returnee high-level subsidy | Employer | Focuses on AI, IT, new energy, new materials, aerospace, biotech, quantum, low-altitude economy | One-off 300k; Chinese, <=50, PhD abroad, returned after 2021 and now full-time; those in other national programs excluded; many places match 1:1 (e.g. Chengdu matches by national/provincial/city tiers) |
Overseas-talent postdoc special | Mobile/workstation | For transitional entry | National + local + unit triple funding |
Boxin program | Mobile/workstation | 2026 born on/after 1 Jan 1994, PhD within 3 years; returnee and foreign PhDs ineligible; in-service may not apply; about 500 funded in 2026 | National 2-yr: 280k/yr + one-off 80k fund; 2026 online 24 Feb to 24 Mar; per National Postdoc Committee current notice, birth line rolls annually |
China Postdoc Science Fund | Mobile/workstation | In-station postdocs | General 80k (natural science, 2026 plan about 300), special 180k (2026 plan about 1,200); full-time in-station, in-service may not apply; per Guide 2026 (中博基字〔2026〕1号) |
RJ one-liner: RJ is a finance-system talent program; industrial-software PhDs mostly need not consider it and may skip directly.
Current-check note (as of Oct 2026, added in revision): the four items - Overseas Outstanding Youth, Boxin, returnee subsidy and Postdoc Science Fund - have funding standards, age lines and windows cross-checked one by one against the 2026 current notices of NSFC, the National Postdoc Committee, the Postdoc Science Fund Council and MOHRSS (incl. host-unit and local HR-bureau forwards, multi-source consistent). The birth dates 1 Jan 1986 and 1 Jan 1994 are the 2026 calibers, rolling annually; next-year applications must follow the current guide, never reuse these values.
(III) Three programs most worth watching (industrial-software specific)
Rank | Program | Why |
1 | QM (enterprise) | Most opportunities are on the enterprise R&D side; QM is enterprise-hosted and need not fully return during application, naturally fitting build-tools roles |
2 | Returnee high-level subsidy | Focuses on IT/IC/high-end-software hardtech, one-off + 1:1 matching, wider threshold (<=50) |
3 | Overseas Outstanding Youth | Highest value for numerical-method/EDA-algorithm academic paths, locks tenure and high matching (must apply before fully returned) |
(IV) Provincial/city talent-program matrix (industrial-software focus)
When choosing a city, treat whether it has industrial-software/EDA/smart-manufacturing talent specials and industry support as a filter - cities with specials have denser windows and more flexible hosts. This matrix now covers about 30 provincial-level/divisions/regions (incl. provincial capitals and top-100 and hotspot cities); each entry lists only representative public program names, no amounts expanded; names, conditions and all amounts are per the current notice of local talent offices/competent authorities.
Prov./city | Representative programs/direction | Fit |
Beijing | High-Innovation Plan Youth Star, Zhongguancun and district talent policies | CAE/EDA/industrial AI, via institutes/new R&D bodies |
Shanghai | Oriental Talent, Pujiang program | EDA/CAE/industrial software, via universities/institutes/Zhangjiang-Lingang |
Nanjing/Suzhou | Double-Innovation talent, Zijinshan talent, industrial-software/IC specials | CAE/EDA/industrial software (strong host) |
Zhejiang/Hangzhou | Kunpeng Action | Industrial AI/digital twin |
Wuhan/Hubei | Chutian Talent, Optics Valley 3551, global-capital support | CAE/industrial software/photonics |
Hunan/Changsha | Furong Plan | CAE/construction-machinery intelligent simulation |
Guangdong/Guangzhou-Shenzhen | Guangdong 15th FYP digital transformation, industrial software; Zhujiang talent plus GZ-SZ overseas parks and funds | Industrial software, industrial AI, entrepreneurship |
Tianjin | Haihe talent, Tiankai science-education-park special; 15th FYP foundational/industrial-software R&D | Industry application software, management software |
Liaoning/Shenyang-Dalian | Xingliao talent, Digital Liaoning smart-province fund, provincial key special (industrial CAE, led by Dalian Univ. of Tech) | CAE solver, industrial-control software (security note) |
Shaanxi/Xian | Sanqin talent, Chuangyuan platform; aerospace and defense simulation scenarios | CAE, industrial OS (security/vetting note) |
Anhui/Hefei | Anhui 15th FYP develops foundational/industrial-software high-end industry; USTC and compute platforms | EDA, industrial AI, algorithms |
Shandong/Jinan-Qingdao | Taishan talent series; 15th FYP breaks through high-end and embedded industrial software, China Software Famous City | High-end industrial software, embedded |
Sichuan/Chengdu-Chongqing | Sichuan 15th FYP targets industrial software; Chongqing strengthens first-version certification; Rongpiao plan | Embedded, industrial OS, automotive software, entrepreneurship |
Heilongjiang/Harbin-Jilin/Changchun | Longjiang science-yingcai and Dingxiang week, Jilin high-level-talent policy; university CS/control backing | Algorithms, embedded, automotive software, postdoc transition |
Fujian/Xiamen-Fuzhou | Fujian software-tech-innovation grants (300/200/100 ten-thousand tiers, per current doc), Xiamen 1:1 returnee matching | Industrial software, domestic-substitution research |
Henan/Zhengzhou-Jiangxi/Nanchang | Zhengzhou Hi-tech-zone software and smart-terminal special; Ganpo talent and aerospace-equipment simulation | Application software, CAE, embedded |
Hebei/Shijiazhuang | Yanzhao Huangjintai Talent-Cluster Plan (backbone talent, tech platform), talent-strong-Hebei; electronic-info and digital-industry research | Industrial software, embedded, domestic-substitution |
Shanxi/Taiyuan | Sanjin Yingcai Plan; Xinchuang and energy-digitalization scenarios | Embedded, industry application software |
Inner Mongolia/Hohhot | Yingcai-Xingmeng talent plan; first-version software incentive (see policy pack) | Industry application software, embedded |
Guangxi/Nannang | Bagui Scholar / Bagui Young Top talent, Guangxi high-level talent (A-E tiers), Young Sprout PhD support; ASEAN digital cooperation | Industry application software, embedded |
Hainan/Haikou | Nanhai Yingcai, Hainan free-trade-port talent (incl. individual-income-tax preference) | Industry application software, digital economy |
Yunnan/Kunmang | Xing-Dian Yingcai plan; South-Asia/Southeast-Asia information cooperation | Industry application software, embedded |
Guizhou/Guiyang | Qianling Yingcai, high-level innovation talent; big-data and compute industry (East-Data-West-Compute hub node) | Industrial AI, compute, embedded |
Gansu/Lanzhou | Longyuan / Western talent plan; renewable-energy and equipment-manufacturing simulation | CAE application, embedded |
Qinghai/Xining | Qinghai-Plateau / Kunlun Yingcai (mostly case-by-case) | Industry application software, embedded |
Ningxia/Yinchuan | Saishang Yingcai plan; national integrated compute-network hub node | Industrial AI, compute, embedded |
Xinjiang/Urumqi | Tianshan Yingcai plan; frontier informatization and energy digitalization | Industry application software, embedded |
Tibet/Lhasa | Zhufeng / Xueyu Yingcai (case-by-case, establishment and title green channel) | Industry application software, frontier informatization |
(V) No-hat roles: landing without a title
Many quality roles release as no-hat: specially-appointed researcher/associate, R&D chief/pre-research PI, tech director, industry professor, enterprise postdoc. They do not require program selection; the employer signs directly with salary/compute terms, then applies via the host to QM-type programs - land first, add later. Fits those not yet meeting Overseas-Youth rigid conditions or whose results are patents/products/kernel code rather than papers.
Type | Note |
Enterprise R&D chief/pre-research PI | No-hat premise, direct signing as kernel/platform R&D lead, later apply to QM/CC via the host |
University specially-appointed researcher/industry professor | Numerical-method/EDA-algorithm direction, tenure + task, links to later Overseas-Youth application |
Tech director/architect | Engineering and industrialization, fits patent/product-result PhDs |
Enterprise postdoc (in-station) | Bound to a localization key-research task, high exit-retention |
(VI) Hidden high-value policy pack (especially valuable for industrial software)
Program | Key point |
First-version software policy | Selection into first-version evaluation grants application/promotion support, directly raising kernel-role value |
High-tech enterprise recognition | Income tax cut from 25% to 15%; industrial-software/EDA firms generally qualify |
Big Fund / local EDA specials | Big Fund III tilts to EDA bottleneck; many localities set EDA procurement/R&D/industrialization specials (amounts per current announcement) |
Overseas-park rent exemption & tax refund | 1-3 yr rent-free, business/tax agency, tax-refund reward |
Startup guarantee loan | Individual up to 500k CNY, enterprise up to 6.0 mn, many places with interest subsidy |
Tax-free car / IIT preference | Within 1 yr of return, buy one domestic/JV car tax-free by CSC certificate; some regions give IIT preference for high-end scarce talent |
Title green channel / CSC certification | Overseas experience can be recognized; CSC certification is a prerequisite for subsidy/hukou/civil-service |
(VII) Industrial-software pre-startup due-diligence list
For industrial-software/EDA/simulation startups, the pitfalls are validation and payment. Check each hardware condition before returning; any shortfall drags the pace:
# | Must-check | Point |
1 | Real industrial POC | Industrial customers are pragmatic; prototypes without on-site validation rarely win orders - finish at least one POC first |
2 | Data-format compatibility & migration cost | Compatible with mainstream toolchains/formats; is customer migration cost controllable |
3 | Bottom-layer licensing & open-source compliance | Are kernel/solver/EDA algorithms, open-source licenses and third-party licenses clear |
4 | Xinchuang and first-version warehousing | Can you enter the Xinchuang catalog / first-version evaluation, deciding procurement landing |
5 | Graded protection / crypto evaluation / industry access | Security or critical-infrastructure scenarios need graded protection, crypto evaluation, industry qualification |
6 | Data export & classification | Compliance and export limits on industrial data |
7 | Payment cycle & warranty deposit | Long accounts receivable and warranty deposits tie up cash - measure it |
8 | Export controls & political vetting | EDA/advanced simulation involve export controls; defense scenarios need political vetting |
(VIII) Application-window timeliness note
Overseas Youth, QM, CC, returnee subsidy and postdoc types all have fixed windows; most provincial/city programs publish annually on a rolling basis. The first wave after landing is often the denser provincial/city specials. Never miss the not-yet-fully-returned precondition of Overseas Youth.
(IX) Basic research vs industrial delivery: review-logic difference
Applications here must read what reviewers want: academic channels value papers and frontier, industry/engineering channels value system delivery and localization-substitution metrics - paper weight is lower than in academic review.
Dimension | Basic research / academic | Industry / engineering delivery |
Core metric | Papers, theoretical innovation, method frontier | System delivery, localization metrics, deliverable engineering results |
Result form | Papers/patents | Product/kernel/solver/platform + industrial-customer validation |
Host | University/institute (Overseas Youth) | Enterprise/central-SOE institute (QM/CC) |
Common mismatch | Applying a paper list to an enterprise kernel role | Using industrial delivery to judge an academic channel |
VII. T-3 months to landing - offer: bundle the seven items together
Offer negotiation for an industrial-software PhD differs from ordinary job-hunting: talking only about monthly salary loses large hidden value, and production factors like kernel/platform/compute resources matter more for long-term growth than cash. Distinguish verbal promises from written contracts.
(I) Seven-item offer breakdown
# | Item | Note |
1 | Base salary / performance | Cash base; fix the ratio and whether the evaluation standard is reasonable and achievable |
2 | Settling / purchase / rent subsidy | Confirm conditions, method (lump/installment), timing and whether pre-tax |
3 | Research-start / R&D conditions | Core academic resource; the industry side checks compute, tool license, real data and prototype conditions |
4 | Program-application support | Most critical - does the unit commit to recommend QM/CC, with coaching, and a fallback if not selected |
5 | Spouse placement / children schooling / medical green channel | Family support; specify schooling tier and age line for dependents |
6 | Hukou / archive / social security / title green channel | Administrative support; specify handling time and recognition process |
7 | Entrepreneurship-specific terms | Equity ratio, term of venue support, IP ownership, exit and industry-fund matching |
(II) Three industry-specific negotiation points
Point | Why critical | How to negotiate |
Kernel/platform participation right | Decides whether you keep building tools rather than being sent to do implementation | Specify in writing the product line/kernel module and role |
Toolchain and compute resources | Solver/EDA/simulation roles depend heavily on license, HPC and real data | Agree on compute quota, software license and data-access scope |
Application-host fallback | Whether the unit hosts QM/CC and a fallback if not selected | Write the host-application promise, coaching and not-selected minimum |
(III) Written negotiation checklist (contract red lines)
All terms must be written into the contract/supplement; verbal promises, chats and emails lack full legal force; settling/research-start/compute are usually tied to a 5-8 year service period and evaluation, and early departure may require proportional restitution - check each restitution/breach clause and keep written confirmation.
Item | Written-confirmation point |
Role track | Explicitly R&D track vs engineering track (the most critical line for industrial software) |
Kernel/product line | Modules, roles and authorship you participate in |
Host-application | The unit commitment to host QM/CC, coaching team and not-selected fallback |
Service period | Years, restitution ratio and evaluation caliber |
Salary structure | Fixed/performance ratio, payment timing and whether pre-tax |
Compute/data | License, HPC quota and industrial-data access rights |
Result ownership | Patent/software-copyright authorship and the boundary of work-invention vs personal results |
(IV) Security / Xinchuang / export-control compliance checklist (key note)
Industrial software often couples with defense, aerospace and energy critical-infrastructure scenarios, so security and control risk is higher than for generic software roles; confirm upfront:
# | Item | Requirement |
1 | Security / political vetting | Defense-simulation and energy-industrial-control posts need vetting; overseas experience must be truthfully stated and de-classification may be extended |
2 | Export controls (ITAR/EAR) | If you touched controlled EDA/advanced-simulation tools overseas, compile a desensitizable results list before returning and declare it on entry |
3 | Data compliance | Roles on real industrial data mind classification and export limits |
4 | Publication freedom | To keep publication freedom, prefer posts without core classified data |
5 | Result ownership | Pre-agree in writing the ownership of work-invention/kernel-code/software-copyright |
VIII. Landing actions: the 3-6-12 month pace
Close the timeline into an executable checklist. This is the landing point of this article download-the-12-month-action-list.
Phase | Key actions |
T-12 months | Finish four-path positioning; verify your institution on the MOE foreign-affairs supervision site; compile papers/patents/kernel code/project-delivery records/recommendation letters; check overseas continuous on-post duration |
T-6 to 9 months | Lock 2-3 cities by industry-fit; list 10-20 target units (institute simulation center / industrial-software vendors / EDA vendors / strong-university institutes); email groups and talent offices to confirm host-application qualification |
T-3 months | Start CSC-certification materials; confirm R&D track and host fallback; negotiate the offer by seven items and check service/restitution clauses |
Landing | Complete CSC certification and onboarding/entry; start the first wave of provincial/city programs; handle hukou, social security, archive and children schooling |
Landing +3 months | First thing: with the unit confirm the national-program (QM/CC/Overseas Youth) contact and timetable, avoiding on-boarded-but-no-one-owns-application |
IX. Zhengxianling®landing point
The biggest barrier for an industrial-software PhD is not knowing there is a gap, but knowing which institute/vendor kernel role truly fits and whether that unit can support applying to QM/CC. Zhengxianling®decomposes and lands these two questions.
To talent: Zhengxianling®serves domestic and overseas PhDs, postdocs, senior R&D engineers and scholars in industrial software and related directions (CAD/CAE/EDA, numerical simulation, industrial AI, industrial OS), offering full-chain support across track positioning, city matching, platform screening, program application, offer negotiation and landing services. The platform keeps open channels for Overseas Outstanding Youth and same-level national programs (QM, HJ, CC, RJ, etc.), synchronizes provincial and municipal talent programs (e.g. Pujiang, Kunpeng, Chutian, Taishan, Kongque, Rongpiao, Zijinshan, Optics Valley 3551, etc.), and hosts via national first-class universities, hospitals, research institutes, central/state SOEs, enterprises, financial institutions, overseas-returnee parks, high-tech zones and some local-government hosts to match academic, enterprise-R&D, entrepreneurship, postdoctoral and no-hat roles. It does not become just another resume-drop box; it aggregates the roles and application channels scattered across research groups, central-SOE institutes, top vendors and local talent fairs into one connectable service line, helping talent clarify item by item who is hiring, whether one can apply through it and how to enter. It is planning to launch an AI multilingual service platform to further reduce cross-border information and language barriers and serve users in all countries; it makes no approval commitment, and final terms follow the employer, competent authority and official current announcements.
To employers/government: universities, research institutes, high-tech zones and industrial-software/EDA/simulation firms needing high-level-talent recruitment, joint-lab building or overseas-top-team project landing may negotiate via the Enterprise Cooperation column of the Zhengxianling®site, forming a role-listing, directed-recommendation, fair-matching and application-coaching loop.
Low-threshold inquiry: industrial-software PhDs/postdocs at a key decision node may download this article 12-month action list and book a 1-on-1 return-path diagnosis. First contact needs only: name, current country/region, highest degree and school, research direction (CAD/CAE/EDA/industrial AI/industrial OS), years of experience and intended region; the platform gives a preliminary matching assessment and operable positioning advice for the sub-direction.
Appendix A: four traps for returning industrial-software PhDs
Trap | Consequence | Fix |
Thinking knowing-to-model/use-software equals building-tools | Applied to a kernel/solver role but sent to application implementation | Self-test by capability labels: have you written kernel operators/solvers/EDA algorithms |
Return first, find a unit later | Overseas Youth requires not-yet-fully-returned, so onboarding forfeits it; the target unit may lack host qualification | Confirm the host and window timetable before returning |
Watch only subsidies, ignore track | High settling fee but placed in the engineering track, limiting bargaining and growth | Distinguish R&D vs engineering track on the first offer line (in writing) |
Ignore security/export controls | Overseas controlled-tool experience and defense scenarios trigger longer vetting or ownership disputes | Run the compliance list before applying and compile a desensitizable results list |
Treating consultancy estimates as official data | Challenged on caliber externally; salary source called inaccurate in negotiation | Tag each industry figure with publisher and caliber (official/agency/media) and keep a verifiable current source |
Appendix B: pre-release verification list and compliance red lines
(I) Pre-release verification list (industrial-software specific)
# | Item to verify | Authority |
1 | 工信厅规〔2024〕33号: by 2027 renew about 2.0 mn industrial-software units and 800k industrial-OS units across 27 key industries | MIIT General Office original (20 Sep 2024); figures and industry list per original |
2 | Localization (overall about 35%; management about 70%, production-control about 52%-60%, R&D-design about 10%-26%; CAD about 20%, CAE and EDA below 10%) | CCID 2025; MIIT 5th Institute 2025; Zhiyan; Industrial Software Alliance current calibers; this guide is an agency-estimate illustration, not official statistics |
3 | Three market-size calibers (MIIT product 333.0 bn / consultancy about 364.97 bn / IDC core 40.88 bn) | MIIT Operations; CCID, Zhongshang and IDC current releases; the three calibers are not comparable |
4 | Talent gap (key-software new gap about 830k, industrial software about 120k; EDA composite is agency-survey) | Key-Software Talent Demand Forecast (2021 forecast) and public research; reference only, no precise citation |
5 | Overseas Youth 1.0-3.0 mn/3yr, born on/after 1 Jan 1986, 36+ months overseas, not returned or returned on/after 1 Jan 2025, full-time in China 3+ years after funding | NSFC current program guide (2026 application closed 15 Jun) |
6 | Boxin 280k/yr (2-yr national) + one-off 80k, born on/after 1 Jan 1994, PhD within 3 years, returnee and foreign PhDs ineligible, in-service may not apply | National Postdoc Committee current notice (Tier A) |
7 | Returnee subsidy one-off 300k, <=50, PhD abroad, returned after 2021 and now full-time; Postdoc Fund general 80k, special 180k (natural science) | MOHRSS and Postdoc Science Fund Council current notices (中博基字〔2026〕1号) |
8 | QM/HJ/CC/RJ conditions and host requirements | MIIT and competent authorities current notices |
9 | Provincial/city industrial-software/EDA subsidies, hukou and settling amounts | Local talent-office / competent-authority current announcements |
10 | First-version evaluation list, Model-Data Resonance covered-industry count | MIIT current announcement / authoritative-agency caliber |
11 | Security / export-control role compliance | Employer and competent-authority terms |
12 | MIIT 2025 Software Operations data (15,483.1 bn; industrial 333.0 bn; foundational 214.6 bn; IC-design 442.1 bn) | MIIT (30 Jan 2026); updates annually, replace with current statistics before citing |
13 | IDC core industrial software and AI+ forecast (40.88 to 84.31 bn, CAGR about 15.6%; AI sub-market CAGR about 49.0%) | IDC China Core Industrial Software and Industrial AI Forecast 2026-2030; agency forecast, note publisher and year |
14 | Provincial 15th-FYP and special docs (Hubei global capital, Liaoning industrial CAE, Jiangsu AI-plus action plan, Model-Data Resonance 20 industries) | Provincial plans and competent-authority public originals; names and statements per current text |
15 | Whether U.S. key-software export controls have extended to industrial software | Per current BIS announcement and entity list; no precise citation without independent verification |
16 | Cross-article consistency of industry data | A fact-layer supplement; before release the platform must write it back to the content baseline as a unified caliber to avoid conflict with sibling articles |
(II) Writing compliance red lines (whole-article self-check)
Red line | Status |
Time caliber unified as the 15th Five-Year (2026-2030) | Kept; unified throughout |
Central SOEs/institutes/vendors all categorized | Kept; a national numerical-simulation center / a leading 3D-CAD vendor / an EDA vendor / an industrial-OS vendor, no real names |
Program codes: QM/HJ/CC/RJ as codes only, no invented full names | Kept; RJ is clearly finance-system and mostly not applicable here, may skip |
All amounts/industry calibers tagged per official and authoritative current | Kept |
Unverifiable data not asserted with a specific figure | Kept; graded as industry caliber, reference only |
Security/export-control tracks add compliance notes | Kept; a dedicated chapter |
Zhengxianling® always with ®; English unified as Zhengxianling® | Kept |
Core phrases
Usage | Phrase |
Cover title | Return of industrial-software PhDs - the window is the bargaining period; the opportunity is at the build-tools kernel. |
Lead sub-title | Many can use software; very few can write out a geometry kernel, a solver or an EDA algorithm line by line. |
Closing action line | Positioning, city, platform, application, offer - turn the five steps into your next-12-month calendar; download the 12-month action list. |
Conclusion
Returning as an industrial-software PhD is not finding a job but seizing a domestic-plus-intelligent window opened by policy, controls and AI together. The value of the timeline method is that it breaks the three questions - why the track matters, where the cities and platforms are, and how far Zhengxianling®can help you land - into your concrete monthly actions for the next 12 months. Remember three prepositions: positioning before applying, platform and host qualification before onboarding, and the application window before landing.
All industry data, policy document numbers, amounts and program calibers in this article are public calibers; final terms follow the official, authoritative-institution, competent-authority and employer current announcements and agreements.
References
Listed by traceability, all public documents or public reports; re-verify against current editions before citing: MIIT General Office Guidelines on Equipment Renewal and Technological Upgrading in Key Industries (工信厅规〔2024〕33号, 20 Sep 2024); MIIT General Office and NDA Comprehensive Dept Notice on the 2026 Model-Data Resonance Action (工信厅联科函〔2026〕193号, 28 Apr 2026); MIIT AI+Software Special Action Implementation Plan (工信部信发〔2026〕209号, issued 2 Sep 2026, press-conf record 11 Sep); MIIT and seven other departments AI+Manufacturing Opinion (Jan 2026); MIIT 2025 Software Industry Operations (30 Jan 2026); SCIO Good Start of the 15th FYP MIIT session (26 Aug 2026); National 15th-FYP Plan; NSFC Overseas Outstanding Youth 2026 Guide; National Postdoc Committee national-funded postdoc program (Tier A Boxin) and China Postdoc Science Fund Guide 2026 (中博基字〔2026〕1号); MOHRSS returnee subsidy and local HR forwards; Key-Software Talent Demand Forecast (2021); public texts of the Hubei, Liaoning, Jiangsu, Shandong, Tianjin, Sichuan, Chongqing, Anhui, Hunan, Guangdong, Henan, Jiangxi, Heilongjiang, Jilin and Fujian 15th FYP, plus the Hubei Action Plan 2024-2026, the Hunan key-software rules (Apr 2026), the Jiangsu AI-plus Industrial-Software Breakthrough Action Plan 2026-2028, Inner Mongolia 2024 policy list item 52 and the Fujian software-tech-innovation grant doc; the Liaoning Dept of Industry and IT proposal reply (2026-06-25); the Wuhan Digital-Intelligence Economy 15th-FYP Plan (Sep 2026); MOE CSC (certification, Chunhui Cup); industry public research and statistics (CCID, IDC, Zhiyan, Industrial Software Alliance, CAICT, Zhongshang, etc.); the national college-employment platform and employer recruitment info. The above industry calibers and agency forecasts follow the original report and official current edition.
About Zhengxianling®
Zhengxianling®is an international high-level-talent platform integrating talent recruitment, science-tech matching, academic linkage and capital services, serving all talent tiers with a current focus on national high-level-talent attraction. For overseas PhDs, postdocs, senior scholars and industry R&D talent it provides channel matching, material pre-screening, policy verification and application coaching; it hosts via national universities, hospitals, research institutes, central/state SOEs, leading firms, financial institutions, overseas-returnee parks, high-tech zones and some local-government hosts, landing provincial/city programs and no-hat roles; it also provides talent sourcing, project matching and think-tank resources to governments, industrial firms and investment institutions. Starting from national talent-application services, it expands to all tiers and going-out services, and plans to launch an AI multilingual service platform to serve users in all countries. The platform makes no approval commitment; policy interpretation and matching suggestions from the AI platform are auxiliary references only, and final terms follow the employer, competent authority and official current announcements.
Contact us
Teacher He (international talent-service consultant)
Email: hb@zhengxianling.com
WhatsApp: +86 189 1352 1933
WeChat: hb798081
WeChat service account: Zhengxianling®
Zhengxianling®website: https://zhengxianling.com
For first contact please provide: name, current country/region, highest degree and graduating school, research direction, years of experience and intended development region (return/come to China), for a preliminary matching assessment.
Positioning, city, platform, application, offer - a five-step decision guide for overseas high-level talent returning to or coming to China.
