广告
加载中

AI4Engineering新锐企业原力引擎获近5000万元天使+轮融资

龚作仁 2026-09-03 17:35
龚作仁 2026/09/03 17:35

邦小白快读

EN
全文速览

本文核心信息是AI4Engineering领域的新锐企业原力引擎完成近5000万元天使+轮融资,以下是核心干货信息:

1. 企业与业务概况:原力引擎由顶尖学术专家吴泰霖教授创立,聚焦面向复杂工程物理系统,构建覆盖仿真、控制、诊断与设计的AI原生底座,以可控核聚变为首个验证场景,能力已拓展至油气、核电、3C、航空航天等能源与高端制造领域。

2. 核心技术优势:对比传统数值求解器,原力引擎的方案可实现数十倍到数百倍的计算加速,能将传统小时、天级的计算压缩至秒级,已经完成聚变装置上机验证和沙特阿美千万级网格油藏仿真工业部署。

3. 本轮融资用途:资金将主要用于技术研发、标杆场景验证、行业客户拓展以及顶尖人才团队建设。

本文给AI硬科技领域品牌商提供了多方面的参考干货,具体如下:

1. 品牌发展路径参考:原力引擎选择以可控核聚变这种对精度、实时性、安全性要求最严苛的高难度场景切入,反向锤炼通用技术能力,再向多行业迁移,这种路径更容易建立高技术壁垒,也更容易获得资本和产业端的认可。

2. 研发方向参考:原力引擎的发展贴合国家政策方向,国务院“人工智能+”行动、“十五五”规划都明确支持AI+工程、核聚变能产业发展,贴合政策方向更容易获得资金、产业资源等支持。

3. 资本合作参考:硬科技品牌打造需要突出自身差异化技术优势,原力引擎兼具学术底蕴和工程落地能力,成功获得多家头部产业资本联合投资,可依托投资方资源快速拓展落地。

本文给面向能源、高端制造领域的技术服务卖家,梳理了行业机会与可借鉴经验,具体如下:

1. 市场机会层面:当前AI正从数字世界走向物理世界,AI4Engineering是全新的增量赛道,政策明确支持产业发展;传统工程领域的仿真、设计环节依赖高成本计算,效率低下,市场对高效率、高精度的AI原生方案需求强烈,存在大量市场空白。

2. 发展路径参考:优先拿下高难度标杆场景完成技术验证,再向通用行业拓展能力,更容易获得客户信任,原力引擎通过聚变场景验证后,已经成功拿下沙特阿美等标杆客户的合作,技术复用性强。

3. 风险提示:AI工程方案必须满足高精度、泛化性、安全边界、现场部署等工程约束,不能只做纯理论产品,需要结合产业实际需求打磨技术,才能落地商业化。

本文面向制造类工厂,梳理了AI赋能生产研发升级的相关干货,具体如下:

1. 研发升级机会:传统工厂在产品研发环节,多物理场仿真、设计优化、参数调试等工作依赖传统CAE软件,面对多耦合复杂系统,计算周期长、成本高,原力引擎的AI原生方案可以将计算从天级、小时级压缩到秒级,能大幅降低研发成本,缩短新品研发周期。

2. 数字化升级方向:AI4Engineering已经可以覆盖工程系统全生命周期的仿真、控制、诊断、设计全环节,工厂推进生产研发数字化升级,可以重点关注这类能够进入真实工程闭环的AI方案,切实提升生产研发效率。

3. 商业合作机会:当前原力引擎等AI技术企业正在向3C、高端制造、航空航天等领域拓展产业合作,有相关研发需求的工厂可以对接这类技术服务商,优化自身研发流程,提升核心竞争力。

本文面向AI技术服务、工业服务类服务商,梳理了行业发展相关干货,具体如下:

1. 行业发展趋势:AI4Engineering作为AI落地物理世界的核心方向,当前已经获得政策和资本的双重推动,随着重大科学装置和先进制造系统加快智能化升级,市场对能够进入工程闭环的AI方案需求快速增长,赛道未来空间广阔。

2. 客户核心痛点:传统数值求解器和CAE软件面对多物理场、多尺度耦合的复杂工程系统,计算成本高、周期长,无法满足参数扫描、反演设计、控制训练等环节的高效率需求,同时客户要求方案满足高精度、安全边界、现场部署等多个工程约束。

3. 技术路径参考:可参考原力引擎的发展路径,以高难度标杆场景打磨技术底层能力,以传统数值和实验数据为基础,结合神经算子、扩散模型等技术实现快速预测,再针对不同行业做适配,打造全链路AI原生底座。

本文面向硬科技产业平台、投资平台,梳理了相关运营和布局干货,具体如下:

1. 赛道企业的核心需求:AI4Engineering领域的硬科技企业,大多从实验室技术转化而来,需要平台提供资本支持、产业资源对接、落地场景匹配等多方面支持,帮助技术从实验室走向真实产业闭环。

2. 平台招商布局方向:AI4Engineering属于国家重点支持的未来产业方向,技术壁垒高,打磨后的技术可复用拓展至多个大市场,这类具备顶尖团队、选择从高难场景切入的硬科技企业,属于优质的招商和投资标的。

3. 平台运营启示:平台可围绕未来产业全产业链做系统性布局,比如围绕聚变能源全产业链布局AI相关核心环节,推动AI和产业深度融合,吸引优质企业落地,构建本地特色未来产业创新生态,同时整合产业资源帮助企业快速落地技术。

本文给AI、工程领域的研究者,梳理了产业发展最新动向和研究参考干货,具体如下:

1. 产业最新动向:AI4Engineering是AI for Science面向产业落地的核心分支,当前已经进入商业化早期阶段,获得了政策和资本的双重支持,行业核心方向是构建贯穿工程系统全生命周期的AI原生闭环,有望重塑复杂工程系统的研发范式。

2. 行业新的发展路径:当前行业共识是,AI4Engineering的突破不在于单点替代传统CAE软件,而在于构建覆盖仿真-控制-诊断-设计全链路的AI原生底座,主流可行路径是选择高难度标杆场景如可控核聚变,反向锤炼技术,再向多行业迁移复用底层能力。

3. 研究落地启示:当前国家重点支持AI+工程、核聚变等未来产业方向,研究者可以结合产业真实需求开展研究,走产学研结合的路径,更容易实现技术突破,也更容易获得资本和产业资源支持,推动成果落地。

返回默认

声明:快读内容全程由AI生成,请注意甄别信息。如您发现问题,请发送邮件至 run@ebrun.com 。

我是 品牌商 卖家 工厂 服务商 平台商 研究者 帮我再读一遍。

Quick Summary

This article covers the key announcement that Origin Force Engine, an emerging player in the AI4Engineering space, has completed a nearly RMB 50 million Series Angel+ funding round. The core takeaways are as follows:

1. Company and business overview: Founded by leading academic Professor Tailin Wu, Origin Force Engine focuses on building an AI-native full-stack infrastructure covering simulation, control, diagnostics and design for complex engineering physics systems. It started with controlled nuclear fusion as its first validation use case, and has now expanded its capabilities to energy and advanced manufacturing sectors including oil and gas, nuclear power, 3C electronics and aerospace.

2. Core technological advantages: Compared with traditional numerical solvers, the company’s solution delivers 10x to 100x computing acceleration, compressing calculation times that traditionally take hours or even days down to seconds. It has completed on-site validation on fusion devices and industrial deployment for a 10-million-grid reservoir simulation project for Saudi Aramco.

3. Use of proceeds: The new funding will primarily go toward technological R&D, benchmark use case validation, industry customer expansion, and recruiting top talent.

This article provides actionable insights for brands in the AI hard tech sector:

1. Brand development path reference: By choosing to enter the highly demanding field of controlled nuclear fusion—where strict requirements for accuracy, real-time performance and safety apply—Origin Force Engine refined its general-purpose technical capabilities before expanding into multiple sectors. This approach makes it easier to build high technical barriers and win recognition from both capital and industry players.

2. R&D direction reference: Origin Force Engine’s growth aligns closely with China’s national policy priorities. Both the State Council’s "AI+" action plan and the 15th Five-Year Plan explicitly support the development of AI plus engineering and the nuclear fusion energy industry. Aligning with policy priorities makes it easier to access funding and industrial resources.

3. Capital partnership reference: Building a hard tech brand requires highlighting differentiated technical advantages. With strong academic foundations and proven engineering delivery capabilities, Origin Force Engine successfully secured co-investment from multiple leading industry investors, and can leverage its investors’ resources to accelerate expansion and commercialization.

This article summarizes industry opportunities and actionable lessons for technology service sellers targeting the energy and advanced manufacturing sectors:

1. Market opportunities: AI is currently expanding from the digital world into the physical world, and AI4Engineering is an entirely new incremental track with clear policy support. Traditional simulation and design processes in engineering rely on high-cost, low-efficiency computing, so there is strong market demand and substantial unmet need for high-efficiency, high-precision AI-native solutions.

2. Development path reference: Securing a technically demanding benchmark use case to validate capabilities before expanding into broader general industries is an effective approach to building customer trust. After validating its technology through fusion use cases, Origin Force Engine successfully landed benchmark clients including Saudi Aramco, thanks to the high reusability of its underlying technology.

3. Risk warning: AI engineering solutions must meet strict engineering requirements for accuracy, generalizability, safety boundaries and on-site deployment. Purely theoretical products will not succeed; technology must be refined in line with actual industrial demands to achieve commercial adoption.

This article outlines key insights for manufacturing factories looking to upgrade R&D and production through AI:

1. R&D upgrade opportunities: For traditional factories, multiphysics simulation, design optimization and parameter tuning in product R&D mostly rely on legacy CAE software. For complex multi-coupled systems, these tools deliver long calculation cycles and high costs. Origin Force Engine’s AI-native solution compresses calculation times that previously took days or hours down to seconds, greatly cutting R&D costs and shortening new product development cycles.

2. Digital upgrade direction: AI4Engineering already covers simulation, control, diagnostics and design across the full lifecycle of engineering systems. Factories pursuing digital upgrades for production and R&D should prioritize AI solutions that can integrate into real-world engineering closed loops to deliver tangible efficiency gains.

3. Business collaboration opportunities: AI technology companies including Origin Force Engine are currently expanding industry partnerships in 3C, advanced manufacturing, aerospace and other sectors. Factories with relevant R&D demands can partner with these technology service providers to optimize their R&D processes and boost core competitiveness.

This article summarizes key industry insights for AI technology service and industrial service providers:

1. Industry development trends: As a core direction for AI’s expansion into the physical world, AI4Engineering is currently backed by both policy and capital. As major scientific facilities and advanced manufacturing systems accelerate their intelligent upgrades, market demand for AI solutions that can integrate into engineering closed loops is growing rapidly, and the track has enormous long-term potential.

2. Core customer pain points: Traditional numerical solvers and CAE software, when applied to complex engineering systems with multi-physics and multi-scale coupling, suffer from high computing costs and long cycles, and cannot meet the high efficiency requirements for parameter scanning, inverse design, control training and other processes. Customers also require solutions to meet multiple engineering constraints including high accuracy, safety boundaries and on-site deployment.

3. Technical path reference: Providers can follow Origin Force Engine’s development approach: refine underlying technical capabilities in technically demanding benchmark use cases, build on traditional numerical and experimental data to deliver fast predictions using technologies such as neural operators and diffusion models, then adapt the solution for different industries to build a full-stack AI-native infrastructure.

This article outlines operational and layout insights for hard tech industry platforms and investment platforms:

1. Core demands of track players: Most hard tech companies in the AI4Engineering space are commercializing laboratory-born technologies, and require platforms to provide capital support, industrial resource connection, and use case matching to help move technologies out of the lab and into real-world engineering closed loops.

2. Platform investment and recruitment direction: AI4Engineering is a future industrial field explicitly supported by the Chinese government, with high technical barriers and widely reusable technology that can expand into multiple large markets. Hard tech companies with top-tier founding teams that enter the market through high-difficulty use cases are high-quality targets for recruitment and investment.

3. Implications for platform operation: Platforms can build systematic layouts across the full industrial chain of future industries. For example, platforms can layout core AI-related segments across the full fusion energy industrial chain to drive deep integration between AI and industry, attract high-quality companies to settle, build a local characteristic future industrial innovation ecosystem, and integrate industrial resources to help companies commercialize technology quickly.

This article summarizes the latest industry developments and research insights for researchers in AI and engineering fields:

1. Latest industry developments: AI4Engineering is a core branch of AI for Science for industrial commercialization, and has now entered the early commercialization stage with support from both policy and capital. The core industry goal is to build an AI-native closed loop across the full lifecycle of engineering systems, which has the potential to reshape the R&D paradigm for complex engineering systems.

2. New industry development path: There is growing industry consensus that breakthroughs in AI4Engineering will not come from replacing single functions of legacy CAE software, but from building an AI-native full-stack infrastructure covering the full simulation-control-diagnosis-design workflow. The broadly proven path is to refine underlying technology by starting with high-difficulty benchmark use cases such as controlled nuclear fusion, then migrate and reuse the underlying capabilities across multiple sectors.

3. Implications for research commercialization: The Chinese government prioritizes future industrial directions including AI+ engineering and nuclear fusion. Researchers that conduct research aligned with real industrial demands and follow the industry-academia-research collaboration path are more likely to achieve technological breakthroughs, access capital and industrial resources, and bring research outcomes to market.

Disclaimer: The "Quick Summary" content is entirely generated by AI. Please exercise discretion when interpreting the information. For issues or corrections, please email run@ebrun.com .

I am a Brand Seller Factory Service Provider Marketplace Seller Researcher Read it again.

近日,AI for Engineering(AI4Engineering)领跑者「原力引擎(上海)智能科技有限公司」(以下简称“原力引擎”)宣布完成天使+轮近5000万元融资。本轮融资由上海国投旗下上海未来产业基金和上海科创集团共同领投,复旦科创加注领投,龙芯创投、成为资本等跟投,光源资本担任独家财务顾问。本轮资金将主要用于生成式仿真与智能控制引擎的持续研发、聚变等标杆场景的装置级验证、能源与高端制造等行业客户拓展,并持续强化全球顶尖复合型人才团队建设。

原力引擎由吴泰霖教授创立,其本科毕业于北京大学物理学院,获麻省理工学院物理学博士学位,后于斯坦福大学计算机系Jure Leskovec教授课题组从事博士后研究;现任西湖大学工学院特聘研究员、助理教授,西湖大学人工智能与科学仿真发现实验室负责人。他长期研究大规模科学仿真、物理系统控制与科学发现,曾与博士生导师Max Tegmark教授提出“理论学习”的新学习范式和以AI物理学家(AI Physicist)为核心的系列方法,并在生成式AI赋能复杂物理系统仿真、控制、设计方面做出一系列领域开创性工作。

在广义的AI4S概念中,AI4Engineering可被视为面向工程物理系统与产业落地的重要分支,它沿袭AI4S在规律学习与复杂系统建模上的能力,并进一步将AI延伸至仿真、状态感知、设计优化和实时控制等环节,最终形成贯穿工程系统全生命周期的闭环交付。原力引擎致力于面向高维、强耦合的复杂工程物理系统,构建覆盖仿真、控制、诊断与设计的AI原生底座。公司以可控核聚变为首个高难度验证场景,并将相关能力拓展至油气、核电、3C及航空航天等能源与高端制造领域。在上述场景中,模型不仅要“算得快”,还必须满足高精度、泛化性、实时性、安全边界和现场部署等工程约束。

从数值求解到AI原生仿真与控制

传统数值求解器和CAE软件是工程研发的重要基础,但面对多物理场、多尺度、多部件耦合系统,通常需要针对具体几何、边界条件与物理过程构建求解流程,当其进入参数扫描、反演设计和控制策略训练的内循环时,计算与工程成本会被进一步放大。

原力引擎并非简单替代传统求解器,而是以数值求解与实验数据为高可信训练基础,通过神经算子、扩散模型等方法学习物理场演化及条件分布,在推理阶段完成快速预测与耦合生成。公司已在流体、油藏和等离子体等复杂物理仿真任务上实现数十倍至数百倍加速,部分任务可由传统的小时或天级计算压缩至秒级,并且可通过真实装置与工业现场验证持续校准模型。

以核聚变验证技术上限,向复杂工程场景迁移

磁约束聚变涉及极端条件下的复杂等离子体系统的高保真仿真、状态估计、闭环控制等算法挑战。其跨越宏观磁流体与微观湍流尺度、涉及芯部—边界多种复杂物理过程耦合,是AI4Engineering最具代表性的高难度场景之一。近年来,学术界已在TCV装置上验证深度强化学习对多种等离子体位形的磁控制,并在DIII-D装置上实现对撕裂不稳定性的主动规避,表明AI正在逐步进入实际聚变实验闭环。

原力引擎在聚变领域正与国内聚变科研机构及头部商业聚变企业推进等离子体仿真、软着陆控制和位形控制合作,公司研发的生成式AI控制算法已完成装置上机实验验证并实现优越控制性能。公司希望以聚变场景对模型精度、实时性和安全性的严苛要求,反向锤炼可复用的物理建模与控制能力。

这种迁移将复用神经算子、生成式建模、安全控制与多智能体协同等底层方法,再针对各行业的几何结构、边界条件、工况数据和安全规范进行适配。创始团队研发的相关算法此前已在沙特阿美的千万级网格油藏仿真中完成工业部署。公司未来也将在核电多物理场仿真、3C力热仿真、航空航天等方向推进产业合作。

原力引擎致力于让AI进入工程闭环,重塑复杂系统研发范式

政策与产业环境亦在AI4Engineering领域形成合力。国务院“人工智能+”行动提出推动人工智能驱动的技术研发、工程实现与产品落地一体化协同,“十五五”规划纲要则将核聚变能列为重点培育的未来产业方向之一。随着重大科学装置与先进制造系统加快智能化升级,能够进入真实工程闭环、并以系统级性能提升衡量价值的AI4Engineering能力,有望成为AI从数字世界走向物理世界的重要基础设施,重塑复杂工程系统的仿真、设计、诊断与控制范式。

从聚变装置中的等离子体控制,到油藏、核电与高端制造中的多物理场仿真,原力引擎所瞄准的是一类共性的工程根本问题:如何让AI在遵循物理规律与安全边界的前提下,更快、更准确地理解、预测并干预复杂系统。

本轮融资完成后,公司将继续以可控核聚变等高难度场景打磨技术上限,推动仿真、控制、诊断与设计能力向包括能源与高端制造在内的更多领域迁移,并逐步沉淀为可复用、可部署、可持续迭代的AI4Engineering基础设施,推动工程研发从依赖高成本计算与经验试错,走向数据、模型与真实系统协同驱动的新范式。

上海未来产业基金表示:“上海未来产业基金将持续围绕聚变能源‘装置-器件-材料-AI4Fusion-应用’全产业链展开系统布局,支持关键技术与核心环节突破。基金高度重视AI技术与聚变产业的深度融合,着力推动“聚变AI原生”能力建设,积极关注并推动AI for Fusion优质稀缺企业在上海落地,赋能在沪聚变装置的智能化升级与产业化进程,助力上海构建具有全球影响力的未来能源创新生态。”

上海科创集团表示:“上海科创集团充分肯定原力引擎吴泰霖教授团队在AI for Engineering领域的前沿突破。公司从可控核聚变等离子体智能控制这一难题出发,其兼具物理机理与数据优势的AI原生引擎大幅提升仿真效率,并已向油藏、航空航天等场景延展。作为长期投资者,我们看好中国硬科技的全球竞争力,将整合产业资源,支持原力引擎强化核心技术,推动核聚变及工业仿真应用早日落地,在全球科技竞争中助力打造AI4E新格局。”

复旦科创表示:“我们看好原力引擎在AI4Engineering方向的技术积累与产业化潜力。公司以可控核聚变、航空航天等高难度场景验证技术边界,构建覆盖仿真、控制、诊断与设计的AI原生能力,有望推动复杂工程系统研发从经验与高成本计算驱动,走向数据、模型与真实系统协同。我们期待与公司共同加速其在能源与高端制造等领域的落地。”

龙芯创投表示:“作为中国首家以生成式AI为底座的AI for Engineering公司,公司率先从可控核聚变这一工程领域最困难的问题切入——核聚变涉及万亿级变量的高维强耦合控制,目前公司已在模拟环境下实现对等离子体不稳定性的毫秒级控制响应,并与核聚变装置头部玩家达成深度合作。我们看好其技术壁垒与落地能力,核聚变这类高维复杂工程场景打磨出的技术具备极强降维复用能力,可拓展至核电、航空航天、高端制造等广阔市场。期待公司持续迭代物理大模型与工程师Agent,打造国产AI工程仿真核心底座,助力我国高端工业与能源科技突破。”

成为资本表示:“成为资本长期深耕前沿技术交叉领域,我们判断,AI4Engineering的下一个突破点不在于单点替代传统CAE软件,而在于构建覆盖‘仿真—控制—诊断’全链路的AI原生底座。原力引擎团队兼具顶尖物理学术底蕴与硬核工程落地能力,选择以可控核聚变这一多物理场强耦合的终极场景作为试金石,率先完成装置级验证。这种‘以最难场景定义技术上限’的路径,与我们‘少而精、重投入、长期陪伴’的投资哲学高度契合。成为资本期待陪伴原力引擎,将AI4Engineering从实验室推向真实产业闭环,重塑复杂工程系统的研发范式,进而见证AI以人类目标为锚点,为理解并安全干预复杂物理系统提供全新的进化路径。”

注:文/龚作仁,文章来源:Laborer,本文为作者独立观点,不代表亿邦动力立场。

文章来源:Laborer

广告
微信
朋友圈

FAQ回顾

AI4Engineering是什么?

AI4Engineering是AI4S面向工程物理系统与产业落地的重要分支,沿袭AI4S在规律学习与复杂系统建模上的能力,进一步延伸至仿真、状态感知、设计优化和实时控制等环节,最终形成贯穿工程系统全生命周期的闭环交付。

原力引擎是做什么的?

原力引擎是AI4Engineering领域领跑企业,面向高维、强耦合的复杂工程物理系统,构建覆盖仿真、控制、诊断与设计的AI原生底座,以可控核聚变为首个高难度验证场景,相关能力可拓展至能源与高端制造等领域。

AI4Engineering有哪些应用场景?

AI4Engineering可应用于可控核聚变等离子体仿真与控制、油气油藏仿真、核电多物理场仿真、3C力热仿真、航空航天系统研发等能源与高端制造相关工程场景,能够显著提升复杂工程系统的研发效率。

原力引擎的AI仿真技术相比传统方案有什么优势?

原力引擎的AI仿真技术以数值求解与实验数据为高可信训练基础,通过神经算子、扩散模型等学习物理场演化规律,在复杂物理仿真任务上可实现数十倍至数百倍加速,部分传统小时级计算可压缩至秒级,还可通过真实场景验证持续校准模型。

这么好看,分享一下?

朋友圈 分享

APP内打开

+1
+1
微信好友 朋友圈 新浪微博 QQ空间
关闭
收藏成功
发送
/140 0