iAISIM Bootstrapped

iAISIM (Wuxianwei) Deep Dive: A Sub-one-year-old Chinese Physics-AI Simulation Start-up Fusing Data and Physics with Multi-architecture Neural Networks into Second-scale Surrogate Simulators, Having Signed an Angel Fund

iAISIM (Beijing) Technology Co., Ltd., founded in October 2025 with Guo Youzhong as legal representative, ships the physics-AI rapid-simulation platform iAISIM. Built on a data-physics dual-driven, multi-architecture neural fusion, it claims 1000x speed-ups and sub-1% error on structural and thermal tasks, and has signed the Dalian Haichuang Qihang angel fund while remaining bootstrapped at the registry.

Executive Summary

iAISIM (Beijing) Technology Co., Ltd., publicly branded in Chinese as Wuxianwei Zhifang and colloquially as iAISIM, is a Chinese physics-AI simulation start-up incorporated on 16 October 2025 and registered in Beijing. Its core product, the iAISIM platform, is an AI rapid-simulation (surrogate) system for computer-aided engineering (CAE). The technical thesis is a data-physics dual-driven architecture: it learns a mapping from design parameters to physical responses from high-fidelity CAE samples, and bakes governing equations and conservation laws into training so that predictions stay physically consistent. The company claims second-scale inference, speed-ups above 1000x versus classical solvers, and sub-1% relative error on structural and thermal tasks, with some cases needing as few as roughly 20 training samples.

Company Overview

The company was established on 16 October 2025 with registered capital of 6.28 million RMB and paid-in capital of 2.186 million RMB. The legal representative, director, manager and finance lead is Guo Youzhong; the supervisor is Zhao Lingjin. Equity is concentrated in two employee/Founder holding partnerships, Beijing Zhannan Huizhi (70%) and Beijing Tunan Huizhi (20%), with individual shareholders Guo Youzhong (4%), Lyu Bing (3%) and Zhao Lingjin (3%). The registered address moved from Chaoyang District to the Beijing Economic-Technological Development Area on 20 October 2026. Public recruitment shows dozens of open roles in Nanjing (Jiangning) and Shanghai (Xuhui), suggesting an active team build-out despite a 2025 annual-report headcount of two. The firm holds several trademarks and software copyrights and has made one outbound investment.

Product and Technology

iAISIM is positioned as a physics-AI multi-physics simulation engine that turns classical CAE into second-scale feedback. Its stack combines computational solid mechanics, FEA, structural optimization and reliability algorithms with deep-learning architectures, especially PINNs and neural operators. The differentiator it advertises is multi-architecture neural fusion: CNNs, RNNs, MLPs and graph neural networks are unified in one training-and-inference pipeline to build high-fidelity surrogates on geometric and mesh tensor representations. A four-step closed loop (submit data, augment, train, deploy) reduces dependence on expert tuning. Ingestion is direct from Ansys and Abaqus mesh and result files; an AI preprocessing stage converts arbitrary 3D geometries into a unified tensor representation; instant inference returns stress, thermal and other 3D fields.

Applications and Case Studies

The company reports deployments across automotive, aerospace, new-energy equipment and advanced manufacturing. Automotive cases include white-body crash, suspension static analysis, anti-collision beam crash, engine mounts and flange/bolt connection nonlinear problems. Aerospace material on the English site describes turbomachinery blades using a NASA Rotor 67 sector (SOLID186 elements) under centrifugal and thermal loading, and a coupled pore-pressure/thermal brain-tissue case. Energy and general-manufacturing samples cited in trade press include wind-tower door openings, metal anti-collision beams, oil-pump housings, rubber mounts and metal cylinder crush. Claimed metrics are speed-ups above 1000x and relative error within 1%, with good results from as few as about 20 samples, though the company notes accuracy depends on dataset selection.

Business Model and Ecosystem

iAISIM is sold as an enterprise simulation tool rather than a pure SaaS: recruitment for pre-sales technical managers in Shanghai shows the go-to-market leans on POC validation, bid documentation and industry workshops in automotive, aerospace and energy equipment. The value narrative targets design-iteration speed-ups above 10x, product-launch-cycle reductions of 30 to 50%, simulation labor-cost cuts above 25% and compute spend reductions of 40 to 50%. No explicit pricing, license model or recurring revenue has been disclosed. The commercial logic is to replace repeated high-fidelity solves with a trained surrogate while classical CAE retains sample generation and final verification.

Competitive Landscape

iAISIM competes with both international and domestic physics-AI and CAE players. It differentiates from data-driven entrants such as Ansys SimAI by adding physics constraints (PINNs) to improve stability on large-deformation and nonlinear cases. Domestically it sits alongside established Chinese CAE names and against broader surrogate-model and generative-design efforts from NVIDIA, Siemens/Altair and specialist start-ups. The advantage is a clear dual-driven thesis and visible automotive deployment; the challenge is that the field is crowded and most rivals have longer histories, larger teams and deeper capital.

Financing and Capital Structure

At the registry the company remains bootstrapped: paid-in capital is 2.186 million RMB against 6.28 million registered, and the shareholder list shows no institutional venture investor. A trade-press report states iAISIM signed the Dalian Haichuang Qihang angel fund, which would be a meaningful endorsement, but that investor does not yet appear in the disclosed equity structure, so the exact form (equity versus cooperation) is unverified. No valuation has been disclosed.

DISCREPANCIES

1. Incorporation versus founding narrative. The company was incorporated on 16 October 2025, while media describe it as built by top algorithm scientists and senior CAE experts; the precise relationship between the named individuals and the legal entity is not documented. Source grade: OFFICIAL (registry) versus MEDIA.

2. Founder identity. Trade press names Wang Xiaohu (vice general manager) as the public face, yet the legal representative is Guo Youzhong; the founders are described in media without named attribution. Source grade: OFFICIAL (registry) versus MEDIA.

3. Financing form. Registered capital is 6.28 million RMB with 2.186 million paid in and no institutional shareholder; a report of signing the Dalian Haichuang Qihang angel fund is not reflected in the equity record. Source grade: MEDIA, unverified.

4. Performance benchmark. The 1000x speed-up and sub-1% error are company-stated from self-reported cases, without an independent third-party benchmark against classical solvers. Source grade: MEDIA, unverified.

5. Headcount. The 2025 annual report lists two employees, whereas dozens of open roles are posted on recruitment sites in Nanjing and Shanghai. Source grade: OFFICIAL versus MEDIA.

6. Industry breadth. The company lists many industries (aerospace, automotive, energy, consumer, marine, electronics), but publicly detailed cases concentrate on automotive structures plus a few aerospace and energy samples. Source grade: OFFICIAL, unverified.

7. Registered address versus ICP filing. The registered address moved to the Beijing ETDA on 20 October 2026, while the ICP filing body address is recorded in Shanghai Minhang, inconsistent with the business-registry address. Source grade: OFFICIAL.

Risks and Diligence Gaps

The primary gap is verification of the performance claim. The 1000x speed-up and sub-1% error are compelling but come from company case materials without a public, third-party benchmark against Ansys or Abaqus on standardized problems. A second gap is the thin operating history: the entity is barely one year old, the registered headcount was two, and the founding team's technical pedigree is described rather than documented. A third gap is capital: despite an angel-fund signing, no institutional equity is visible and no valuation is disclosed. A fourth gap is competitive pressure in a crowded physics-AI CAE field where incumbents have far deeper resources.

Overall Assessment

  • techScore 7.0. The data-physics dual-driven thesis and multi-architecture fusion are differentiated, and visible automotive cases show real deployment. Deducted points: no independent benchmark, large-deformation stability unproven publicly, thin documented pedigree.
  • bizScore 5.5. Active recruitment and POC-led go-to-market indicate commercial intent, but no pricing, license model or recurring revenue is disclosed. Deducted points: no disclosed revenue, model not public.
  • fundScore 5.0. Paid-in capital of 2.186 million RMB with no visible institutional shareholder; an angel-fund signing is unverified in equity. Deducted points: small capital base, no disclosed valuation, financing form uncertain.
  • productScore 6.5. iAISIM is shipped and demonstrated across multiple industries with a concrete four-step loop and Ansys/Abaqus ingestion. Deducted points: no public benchmark, adaptive meshing still in development, cases concentrated in automotive.
  • marketScore 6.0. The domestic physics-AI CAE opportunity is large and the automotive wedge is compelling, but the field is crowded and incumbents are better funded. Deducted points: young company, limited proven scale, strong competition.

The composite view is 6.0. iAISIM is a very early but visibly active Chinese physics-AI CAE entrant with a coherent dual-driven thesis and real automotive deployment signals, yet it carries the standard early-stage deficits of unverified performance, undocumented pedigree, thin capital and a crowded competitive set. The most informative next disclosures will be an independent benchmark against classical solvers, a first named enterprise reference, and clarity on the angel-fund equity.

References

  1. iAISIM official site (iaisim.com) and English site (iaisim.com/en), 2026: product, technology stack, roadmap and public case studies (NASA Rotor 67, brain-tissue coupling).
  2. iAISIM WeChat introduction, 2026: data-physics dual-driven architecture, four-step loop, industry coverage.
  3. CCID Net, 23 May 2026: iAISIM launched at the 22nd China CAE Engineering Analysis Annual Conference; surrogate model, sub-1% error, value metrics.
  4. China Industry News (cinn.cn), 28 October 2026: iAISIM at SAECCE 2026 in Suzhou; automotive CAE cases (white-body crash, suspension, anti-collision beam).
  5. China Development Net, June 2026: iAISIM at the 8th Simulation Technology Application Conference in Wenchang.
  6. China Machinery Industry Federation (cmif.org.cn), May 2026: iAISIM product showcase and cases (wind tower, anti-collision beam, oil-pump housing).
  7. Fangzhenxiu, 2026: iAISIM builds 1% accuracy second-scale surrogate models; minimum about 20 samples; Ansys/Abaqus ingestion.
  8. China Business News (zgswcn.com), 2026: iAISIM signed the Dalian Haichuang Qihang angel fund.
  9. Company registries (Qichacha, Tianyancha, Qibook), 2025 to 2026: incorporation 16 October 2025, capital 6.28M/2.186M RMB, legal representative Guo Youzhong, shareholders, IP and recruitment.
  10. Technical background on neural operators (FNO, DeepONet) and PINNs underpinning iAISIM's architecture, 2026.
This report is compiled from public sources and does not constitute investment advice. The company is private with no audited financials; the 1000x speed-up, sub-1% error, angel-fund signing, founder pedigree and case details are drawn from company statements and trade press with no independent benchmark, and no revenue or valuation has been disclosed; uncertainties are flagged throughout. Data is current as of 7 November 2026.

Related tools

This report is compiled from public sources. It is for reference only and does not constitute investment advice. Data is current as of the publication date; please refer to the company’s official filings.