广告
加载中

星源智荣获第十一届“创客北京2026”创新创业大赛双项一等奖

龚作仁 2026-10-21 09:41
龚作仁 2026/10/21 09:41

邦小白快读

EN
全文速览

总:星源智在第十一届创客北京2026创新创业大赛中拿下北京市总决赛企业组一等奖和北汽专项赛一等奖,核心看点是具身大模型与机器人通用大脑。

1.大赛数据:本届4月启动、9月收官,共征集5518个项目,人工智能、高端制造、医药健康、能源材料及环保等重点领域项目占比超过五成。

2.技术干货:星源智由北京智源人工智能研究院孵化,专注跨平台、跨本体的机器人通用大脑;具身大脑RoboBrain Pro可一次训练、跨本体复用;世界模型ω‑EVA在LIBERO测试集任务成功率98.6%;端侧平台降低部署成本。

3.落地场景:已用于工业制造、生产分拣、物流搬运、商业服务、教育教学、市政巡检,支持轮式双臂、叉车、人形机器人及搭载机械臂的机器狗。

总:对品牌商而言,星源智的看点主要是技术品牌背书、产品研发体系和生态协同能力。

1.品牌背书:获创客北京2026北京市总决赛企业组一等奖、北汽专项赛一等奖,从5518个项目中脱颖而出,可用于强化技术领先与产业认可形象。

2.研发定位:坚持不做机器人本体,专注跨平台、跨本体的机器人通用大脑,形成具身大脑、世界模型、端侧算力、跨本体调度平台的组合,有利于建立差异化技术品牌。

3.场景与生态:在工业制造、生产分拣、物流搬运、商业服务、教育教学、市政巡检部署,牵头成立具身Tier 1联盟,联合机械臂、轮式底盘、传感器等企业,为整机企业缩短开发周期,可作为B2B品牌合作与生态背书素材。

总:卖家可重点关注政策风向、市场机会和可复制的合作模式。

1.政策与赛事:北京市持续推动科技创新和产业创新融合,完善政策供给、要素保障和生态;本届大赛设区域赛、龙头企业专项赛、OPC专项赛,征集5518个项目,重点领域占比超过五成。

2.机会提示:北汽集团等13家龙头企业张榜出题,中小企业揭榜答题,十余个项目已与链主企业达成合作意向,说明具身智能和产业链协同存在订单与合作入口。

3.可学习点:星源智不做本体,做通用大脑、端侧平台和RoboXross调度平台,并通过具身Tier 1联盟联合零部件企业,降低整机开发周期;卖家可关注其跨本体复用、多场景部署和联盟合作方式。

总:工厂可关注机器人通用大脑带来的生产端降本、部署和多机协同机会。

1.生产场景:星源智已在工业制造、生产分拣、物流搬运、市政巡检等场景部署,支持轮式双臂、叉车、人形机器人及搭载机械臂的机器狗,说明工厂自动化改造可对接多形态设备。

2.产品与设计需求:RoboBrain Pro提供感知、决策、任务规划、时空认知、语音理解表达与推理能力,一次训练可跨本体复用;端侧智能计算平台以软硬一体方案降低机器人本体公司的部署门槛与落地成本。

3.商业机会:公司牵头具身Tier1联盟,联合机械臂、轮式底盘、传感器等零部件企业,为整机企业缩短开发周期;工厂可关注与整机企业、零部件企业及调度平台的协同接入,推进多机器人统一运营和数字化升级。

总:服务商可关注具身智能进入生产现场的趋势、客户痛点和平台化解决方案。

1.趋势:星源智以具身大模型和机器人通用大脑推进物理AGI,赛事和北京政策支持大中小企业融通发展,具身智能正从研发走向工业制造、分拣、物流、商业服务、教育及巡检。

2.客户痛点:行业存在一机一脑局限、机器人本体部署门槛与落地成本高、多形态多机器人统一运营难、任务成功率待提升等问题。

3.解决方案:RoboBrain Pro实现跨本体复用;世界模型ω‑EVA以预演-验证-行动闭环预判环境变化,LIBERO任务成功率98.6%;端侧智能计算平台降低部署成本;RoboXross覆盖接入管理、本体匹配、任务调度、运行监控、事件识别、工单闭环与数据分析。

总:平台商可关注RoboXross这类跨本体调度平台的运营逻辑和生态协同需求。

1.平台需求:多形态机器人接入后,需要统一运营,解决本体匹配、任务调度、运行监控、事件识别、工单闭环与数据分析等全流程管理问题。

2.最新做法:星源智以通用大脑为底座,推出RoboXross跨本体具身大脑调度平台,实现多形态机器人接入和多机器人统一运营;同时通过具身Tier1联盟联合机械臂、轮式底盘、传感器企业,为整机企业缩短开发周期。

3.运营与风向:大赛推动北汽集团等13家龙头企业张榜出题,十余个项目达成合作意向,平台商可关注以真实产业需求连接创新项目的模式;在场景上覆盖工业制造、生产分拣、物流搬运、商业服务、教育教学及市政巡检,平台建设需围绕接入、调度、监控和闭环服务展开。

总:研究者可关注具身智能的技术路线、产业组织方式和政策启示。

1.技术新动向:星源智由北京智源人工智能研究院孵化,围绕具身大脑、世界模型、端侧算力及跨本体调度平台构建物理AGI路径;RoboBrain Pro打破一机一脑局限,一次训练跨本体复用;ω‑EVA首创预演-验证-行动闭环,LIBERO成功率98.6%。

2.产业新问题与模式:机器人本体部署门槛高、多机器人协同难;企业选择不做本体、做通用大脑和调度平台,并通过具身Tier1联盟联合零部件企业,形成平台化、生态化商业模式。

3.政策与赛事启示:北京推动科技创新与产业创新融合,大赛以龙头企业张榜、中小企业揭榜促进融通,13家龙头企业参与,十余个项目达成合作意向,为具身智能产业政策和中小企业专精特新发展提供案例。

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

Summary: Xingyuanzhi won the enterprise group first prize at the Beijing final of the 11th Maker Beijing 2026 Innovation and Entrepreneurship Competition and the first prize in the BAIC special track. The core highlight is its embodied large model and general-purpose robot brain.

1. Competition data: This year's event launched in April and concluded in September, collecting 5,518 projects, with key sectors such as AI, advanced manufacturing, healthcare, energy materials, and environmental protection accounting for more than half.

2. Technology: Incubated by the Beijing Academy of Artificial Intelligence (BAAI), Xingyuanzhi focuses on a general-purpose robot brain that is cross-platform and cross-embodiment. Its embodied brain RoboBrain Pro enables one-time training and cross-embodiment reuse; its world model ω-EVA achieves a 98.6% task success rate on the LIBERO benchmark; and its on-device platform lowers deployment costs.

3. Applications: It has been deployed in industrial manufacturing, production sorting, logistics handling, commercial services, education and teaching, and municipal inspection, supporting wheeled dual-arm robots, forklifts, humanoid robots, and robot dogs equipped with robotic arms.

Summary: For brands, Xingyuanzhi's main appeal lies in its technology brand endorsement, product R&D system, and ecosystem collaboration capabilities.

1. Brand endorsement: It won first prize in the enterprise group at the Beijing final of Maker Beijing 2026 and first prize in the BAIC special track, standing out from 5,518 projects. This can reinforce an image of technological leadership and industry recognition.

2. R&D positioning: It deliberately does not build robot bodies and instead focuses on a cross-platform, cross-embodiment general-purpose robot brain. The combination of embodied brain, world model, on-device computing, and cross-embodiment scheduling platform helps build a differentiated technology brand.

3. Scenarios and ecosystem: It has deployed in industrial manufacturing, production sorting, logistics handling, commercial services, education and teaching, and municipal inspection, and led the establishment of the Embodied Tier 1 Alliance with companies in robotic arms, wheeled chassis, and sensors to shorten development cycles for robot OEMs. This can serve as material for B2B brand partnerships and ecosystem endorsement.

Summary: Sellers should focus on policy direction, market opportunities, and replicable cooperation models.

1. Policy and competition: Beijing continues to promote the integration of technological and industrial innovation and improve policy supply, factor guarantees, and the ecosystem. This year's competition featured regional tracks, leading-enterprise special tracks, and an OPC special track, collecting 5,518 projects, with key sectors accounting for more than half.

2. Opportunity signals: 13 leading enterprises including BAIC Group issued open challenges, and SMEs responded with solutions; more than 10 projects have reached cooperation intentions with chain leader enterprises, indicating order and cooperation entry points in embodied intelligence and industrial chain collaboration.

3. Takeaways: Xingyuanzhi does not build robot bodies but focuses on a general-purpose brain, on-device platform, and RoboXross scheduling platform, and works with component companies through the Embodied Tier 1 Alliance to reduce robot development cycles. Sellers can watch its cross-embodiment reuse, multi-scenario deployment, and alliance-based cooperation approach.

Summary: Factories should focus on production-side cost reduction, deployment, and multi-robot collaboration opportunities brought by a general-purpose robot brain.

1. Production scenarios: Xingyuanzhi has deployed in industrial manufacturing, production sorting, logistics handling, and municipal inspection, supporting wheeled dual-arm robots, forklifts, humanoid robots, and robot dogs equipped with robotic arms, indicating that factory automation upgrades can connect with multiple equipment forms.

2. Product and design needs: RoboBrain Pro provides perception, decision-making, task planning, spatiotemporal cognition, speech understanding and expression, and reasoning, with one-time training enabling cross-embodiment reuse. Its on-device intelligent computing platform uses an integrated hardware-software approach to lower deployment thresholds and implementation costs for robot OEMs.

3. Business opportunities: The company leads the Embodied Tier 1 Alliance with component companies in robotic arms, wheeled chassis, and sensors to shorten development cycles for robot OEMs. Factories can focus on integration with robot OEMs, component companies, and scheduling platforms to advance unified multi-robot operations and digital upgrading.

Summary: Service providers should watch the trend of embodied intelligence entering production sites, customer pain points, and platform-based solutions.

1. Trend: Xingyuanzhi is advancing physical AGI through embodied large models and a general-purpose robot brain. The competition and Beijing policy support integrated development of large and small enterprises, and embodied intelligence is moving from R&D into industrial manufacturing, sorting, logistics, commercial services, education, and inspection.

2. Customer pain points: The industry faces limitations of one brain per robot, high deployment thresholds and implementation costs for robot bodies, difficulties in unified operation of multiple robot forms, and the need to improve task success rates.

3. Solutions: RoboBrain Pro enables cross-embodiment reuse; the world model ω-EVA predicts environmental changes through a rehearsal-validation-action loop and achieves a 98.6% success rate on LIBERO; the on-device intelligent computing platform lowers deployment costs; and RoboXross covers access management, embodiment matching, task scheduling, operation monitoring, event recognition, work-order closed loop, and data analysis.

Summary: Platform operators should focus on the operating logic and ecosystem collaboration needs of cross-embodiment scheduling platforms such as RoboXross.

1. Platform needs: After multiple robot forms are connected, unified operations are required to address full-process management issues such as embodiment matching, task scheduling, operation monitoring, event recognition, work-order closed loop, and data analysis.

2. Latest approach: With its general-purpose brain as the foundation, Xingyuanzhi launched RoboXross, a cross-embodiment embodied brain scheduling platform, enabling access by multiple robot forms and unified operation of multiple robots. At the same time, through the Embodied Tier 1 Alliance, it works with robotic arm, wheeled chassis, and sensor companies to shorten development cycles for robot OEMs.

3. Operations and direction: The competition encouraged 13 leading enterprises including BAIC Group to issue open challenges, and more than 10 projects reached cooperation intentions. Platform operators can watch models that connect innovative projects with real industrial demand. In terms of scenarios, it covers industrial manufacturing, production sorting, logistics handling, commercial services, education and teaching, and municipal inspection. Platform building should center on access, scheduling, monitoring, and closed-loop services.

Summary: Researchers should focus on the technology roadmap, industrial organization, and policy implications of embodied intelligence.

1. New technology directions: Incubated by BAAI, Xingyuanzhi is building a physical AGI path around an embodied brain, world model, on-device computing, and cross-embodiment scheduling platform. RoboBrain Pro breaks the one-brain-per-robot limitation and enables one-time training with cross-embodiment reuse. ω-EVA pioneers a rehearsal-validation-action loop and achieves a 98.6% success rate on LIBERO.

2. New industrial issues and models: Robot body deployment thresholds are high, and multi-robot collaboration is difficult. The company chooses not to build robot bodies but to build a general-purpose brain and scheduling platform, and works with component companies through the Embodied Tier 1 Alliance, forming a platform-based, ecosystem-oriented business model.

3. Policy and competition implications: Beijing promotes the integration of technological and industrial innovation. The competition uses open challenges from leading enterprises and responses from SMEs to promote integration, with 13 leading enterprises participating and more than 10 projects reaching cooperation intentions, providing a case for embodied intelligence industry policy and the specialized, refined, distinctive, and innovative development of SMEs.

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.

第十一届“创客北京2026”创新创业大赛总结会于数智北京创新中心(展示中心)举行。星源智具身大模型项目从5000余个参赛项目中脱颖而出,分别荣获北京市总决赛企业组一等奖、北汽专项赛一等奖。CEO刘东作为获奖企业代表在大会上发言,并介绍公司在机器人通用大脑、世界模型以及产业落地应用方面的进展。

图片

图:<左:北汽专项赛一等奖>;<右:创客北京北京市总决赛企业组一等奖>

图片

本届大赛以“创客北京 共赢未来”为主题,于今年4月启动、9月收官,设置区域赛、龙头企业专项赛、OPC专项赛三大赛道,共征集项目5518个。人工智能、高端制造、医药健康、能源材料及环保等重点领域项目占比超过五成。

此次荣获“创客北京2026”双项一等奖,为星源智进一步拓展产业交流与合作提供了契机。

图片

图片

图:双项荣誉颁奖现场

北京市经济和信息化局党组成员、副局长苏国斌在总结会致辞中指出,北京市持续推动科技创新和产业创新深度融合,不断完善政策供给、强化要素保障、优化发展生态,引导中小企业走专精特新发展道路。

推动大中小企业融通发展,是本届大赛的重要特点。北汽集团等13家龙头企业“张榜出题”,中小企业“揭榜答题”,以实际产业需求连接创新项目,促进产业链上下游协同。大赛期间,十余个项目已与链主企业达成合作意向。

图片

图:北京市经济和信息化局党组成员、副局长苏国斌致辞

这一导向与星源智的技术研发和产品实践相契合。围绕具身智能进入生产现场的实际需求,星源智持续推进通用大脑、世界模型研发、端侧部署和多机器人协同,让模型能力转化为机器人能够执行的具体任务。

图片

“我们坚持不做机器人本体,专注于做跨平台、跨本体的机器人通用大脑。”刘东在发言中介绍。

图片

图:星源智创始人&CEO刘东发言现场

星源智由北京智源人工智能研究院深度孵化,是北京市重点培育的具身智能企业,致力于构建通往物理AGI的具身大模型。公司围绕具身大脑、世界模型、端侧算力及跨本体具身大脑调度平台四个方面,持续完善技术与产品体系。

具身大脑RoboBrain Pro,作为机器人通用智能底座,实现智能能力标准化输出,具备完整的感知、决策、任务规划、时空认知、语音理解表达与推理能力,打破“一机一脑”局限,通过一次训练,实现跨本体复用。

世界模型ω‑EVA,首创“预演-验证-行动”的思维闭环,让机器人在实操前就可以预判动作带来的环境变化,再根据推演结果优化动作方案,大幅提升任务成功率,在LIBERO测试集上任务成功率高达98.6%。

端侧智能计算平台提供高性能软硬一体的方案,降低机器人本体公司的部署门槛与落地成本,把高阶智能赋予机器人本体,推动机器人技术从实验室走向规模化商业应用。

RoboXross跨本体具身大脑调度平台,可实现多形态机器人接入、多机器人统一运营,覆盖接入管理、本体匹配、任务调度、运行监控、事件识别、工单闭环与数据分析全流程。

从理解指令、规划任务,到执行动作、协同作业,星源智正通过通用的大脑底座,支持不同机器人适配多样化的应用需求。

图片

技术的应用价值,需要在实际任务中检验。目前,星源智已在工业制造、生产分拣、物流搬运、商业服务、教育教学及市政巡检等场景开展部署,支持轮式双臂、叉车、人形机器人及搭载机械臂的机器狗等不同形态设备的应用。

在推进产品落地的同时,星源智牵头成立具身Tier 1联盟,联合机械臂、轮式底盘、传感器等领域的零部件企业,推动产业资源协同,为整机企业缩短机器人开发周期提供支持。

从通用大脑研发到场景部署,从单机任务执行到多机协同,星源智持续围绕产业需求完善产品能力,推动具身智能技术与生产环节衔接。

星源智将持续深化具身大模型及通用大脑研发,携手产业链合作伙伴推进场景应用,让机器人在更多生产任务中发挥作用,为北京具身智能产业发展贡献力量。

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

星源智是一家做什么的公司?

星源智是由北京智源人工智能研究院深度孵化的具身智能企业,北京市重点培育的具身智能企业,致力于构建通往物理AGI的具身大模型。公司不做机器人本体,专注于跨平台、跨本体的机器人通用大脑,并布局具身大脑、世界模型、端侧算力及跨本体调度平台。

星源智的具身大模型产品体系包括哪些?

星源智围绕具身大脑、世界模型、端侧算力及跨本体具身大脑调度平台构建产品体系,包括机器人通用智能底座RoboBrain Pro、世界模型ω-EVA、端侧智能计算平台和RoboXross跨本体具身大脑调度平台,覆盖感知、决策、任务规划、多机协同与运营管理。

星源智的机器人通用大脑可以应用在哪些场景?

星源智已在工业制造、生产分拣、物流搬运、商业服务、教育教学及市政巡检等场景开展部署,支持轮式双臂、叉车、人形机器人及搭载机械臂的机器狗等不同形态设备,通过通用大脑底座适配多样化应用需求。

星源智世界模型ω-EVA有什么技术特点?

ω-EVA首创“预演-验证-行动”的思维闭环,让机器人在实操前预判动作带来的环境变化,再根据推演结果优化动作方案,提升任务成功率;在LIBERO测试集上任务成功率高达98.6%。

星源智的RoboBrain Pro如何实现跨本体复用?

RoboBrain Pro作为机器人通用智能底座,实现智能能力标准化输出,具备感知、决策、任务规划、时空认知、语音理解表达与推理能力,打破“一机一脑”局限,通过一次训练实现跨本体复用,降低机器人部署门槛。

这么好看,分享一下?

朋友圈 分享

APP内打开

赞 +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