星海图完成超2亿元Pre-A轮融资｜产业融资快报

邦小白快读

星海图完成超2亿元Pre-A轮融资，在具身智能领域实现关键技术突破和商业化应用。

1.融资事件：由高瓴创投、蚂蚁集团领投，米哈游、无锡创投集团等跟投，华兴资本担任独家财务顾问，金额超2亿元，体现市场认可。

2.技术亮点：公司开发具身基础模型EFM-1，采用模仿学习和3D扩散策略DP3，仅需40次演示数据即可完成操作任务，成功率高达85%，具有强泛化能力。

3.空间智能引擎：全球首个Real2Sim2Real引擎，能基于消费级2D相机重建万平米场景几何，解决数据匮乏问题，并生成逼真渲染图。

4.硬件产品：全尺寸轮式双臂具身本体R1，能执行搬箱子、开关门等全场景任务，适用于多种地形和空间。

5.应用前景：已在智能制造、仓储物流、跨境电商、家用场景探索应用，通过上下游联合共创实现快速落地，显示商业潜力。

星海图的融资和技术进展为品牌营销和产品研发提供新机遇，反映消费趋势变化。

1.品牌营销：知名机构如高瓴创投、蚂蚁集团领投融资，提升品牌可信度，可用于渠道建设和市场定位。

2.产品研发：具身基础模型EFM-1使用创新算法模仿学习，支持双臂操作任务，为智能产品开发提供技术基础。

3.消费趋势：应用在跨境电商、家用场景中，体现用户行为向自动化转型，品牌可据此设计定价策略。

4.用户行为观察：硬件本体R1能完成日常任务如搬箱子，显示消费者对智能生活需求的增长，品牌可研发相关产品。

5.市场前景：投资人如蚂蚁集团强调具身智能市场空间巨大，未来十年潜在价值数十万亿，品牌可利用此趋势拓展商机。

星海图的融资和商业化进程为卖家提供增长市场、机会提示和应对措施。

1.增长市场：应用场景包括跨境电商、仓储物流等，显示新兴需求变化，卖家可探索合作或进入新领域。

2.机会提示：公司通过上下游联合共创实现快速落地，卖家可学习商业模式如部署路径，创造合作机会。

3.正面影响：融资事件和政策支持体现行业潜力，风险提示：早期阶段数据依赖，但Real2Sim2Real引擎解决数据匮乏问题，降低运营风险。

4.事件应对：模仿学习算法仅需40次数据完成任务，卖家可借鉴提升效率；硬件R1操作任务如捡零件，提供实用学习点。

5.扶持政策：投资人如华兴资本提供财务顾问支持，卖家可寻求类似资源，规避风向。

星海图的技术在智能制造等场景提供生产需求启示和商业机会。

1.生产需求：应用在智能制造、仓储物流领域，工厂可借鉴执行搬箱子、捡零件等任务，优化生产线设计。

2.商业机会：公司探索上下游联合共创，工厂可参与合作，如在产品设计或数字化推进中提供支持。

3.数字化启示：空间智能引擎Real2Sim2Real能重建场景几何并解决数据问题，启示工厂如何高效管理大规模数据。

4.技术亮点：具身基础模型EFM-1的泛化能力，可用于定制化生产；模仿学习算法减少数据需求，提升效率。

5.电商推进：在跨境电商场景应用，显示电商化趋势，工厂可借此拓展销售渠道。

星海图的行业趋势和新技术为服务商提供客户痛点解决方案。

1.行业趋势：具身智能是AI物理应用热点，市场空间巨大，投资人如蚂蚁集团预测未来十年数十万亿蓝海市场。

2.新技术：具身基础模型EFM-1和空间智能引擎Real2Sim2Real，突破性解决数据匮乏问题，服务商可集成或开发类似方案。

3.客户痛点：Real2Sim2Real引擎基于一条数据扩展数千条，直接应对数据短缺痛点，提供高效解决方案。

4.创新方向：模仿学习算法仅需少量数据完成任务，服务商可借鉴优化服务流程；硬件R1全场景能力展示应用潜力。

5.发展前景：公司商业化在仓储物流等领域落地，服务商可关注趋势，提供定制化服务。

星海图的商业需求和平台做法为平台商提供招商、运营和风险规避参考。

1.平台需求：应用在跨境电商、仓储物流等场景，显示对平台服务的需求，平台商可招商合作。

2.平台做法：公司通过上下游联合共创和清晰部署路径实现快速落地，平台商可学习运营管理方式。

3.招商机会：融资事件吸引多家机构，平台商可借此吸引类似企业入驻；硬件R1操作任务如开关门，提供合作点。

4.风险规避：Real2Sim2Real引擎解决数据匮乏问题，降低平台运营风险；投资人观点提示早期阶段数据依赖，需注意。

5.效率提升：模仿学习算法的高成功率，启示平台如何优化流程，提升服务速度。

星海图在具身智能领域带来产业新动向、问题解决和商业模式启示。

1.产业动向：开发具身基础模型EFM-1和全球首个空间智能引擎Real2Sim2Real，突破行业标准，显示技术演进。

2.新问题：Real2Sim2Real引擎解决数据匮乏问题，提供研究启示；模仿学习算法仅需40次数据，挑战传统方法。

3.政策建议：投资人如高瓴创投强调数据积累和算法迭代关键，暗示政策支持方向；社会价值巨大，需法规引导。

4.商业模式：探索智能制造、家用等多场景应用，通过联合共创实现快速商业化，研究者可分析路径。

5.创新细节：3D扩散策略DP3在真实任务中泛化能力强，硬件R1全地形能力，为研究提供案例。

Star Atlas has completed a Pre-A Series funding round exceeding 200 million yuan, achieving key technological breakthroughs and commercial applications in embodied intelligence.

1. Funding Event: Led by Hillhouse Capital and Ant Group, with participation from miHoYo and Wuxi Venture Capital Group, and Huaxing Capital acting as the exclusive financial advisor. The over 200 million yuan investment reflects strong market recognition.

2. Technical Highlights: The company developed the Embodied Foundation Model EFM-1, utilizing imitation learning and the 3D diffusion strategy DP3. It requires only 40 demonstrations to complete operational tasks with an 85% success rate and demonstrates strong generalization capabilities.

3. Spatial Intelligence Engine: The world's first Real2Sim2Real engine can reconstruct the geometry of 10,000-square-meter scenes using consumer-grade 2D cameras, addressing data scarcity and generating photorealistic renderings.

4. Hardware Product: The full-scale wheeled dual-arm embodied robot R1 can perform full-scene tasks like moving boxes and opening/closing doors, adaptable to various terrains and spaces.

5. Application Prospects: Applications are being explored in smart manufacturing, warehousing/logistics, cross-border e-commerce, and home scenarios. Rapid commercialization through upstream/downstream co-creation demonstrates significant business potential.

Star Atlas's funding and technological progress offer new opportunities for brand marketing and product R&D, reflecting shifts in consumer trends.

1. Brand Marketing: Leading investment from prestigious institutions like Hillhouse Capital and Ant Group enhances brand credibility, useful for channel development and market positioning.

2. Product R&D: The EFM-1 embodied foundation model uses innovative imitation learning algorithms to support dual-arm operations, providing a technical foundation for smart product development.

3. Consumer Trends: Applications in cross-border e-commerce and home scenarios indicate a shift towards automation in user behavior, informing brand pricing strategies.

4. User Behavior Observation: The R1 hardware's ability to perform daily tasks like moving boxes shows growing consumer demand for smart living, prompting brands to develop related products.

5. Market Outlook: Investors like Ant Group emphasize the massive potential of the embodied AI market, projected to be worth tens of trillions over the next decade, offering brands opportunities for expansion.

Star Atlas's funding and commercialization progress provide sellers with insights into growth markets, opportunities, and strategic responses.

1. Growth Markets: Applications in cross-border e-commerce and warehousing/logistics reveal emerging demand shifts, suggesting sellers explore partnerships or new market entry.

2. Opportunity Indicators: The company's rapid implementation through upstream/downstream co-creation offers a learnable business model, such as deployment pathways, for creating collaboration opportunities.

3. Positive Impact & Risk Note: The funding event and policy support highlight industry potential. A noted risk is early-stage data dependency, but the Real2Sim2Real engine mitigates data scarcity, reducing operational risks.

4. Strategic Response: The imitation learning algorithm's efficiency (requiring only 40 data points) can be adopted by sellers to improve operations; the R1's task execution (e.g., picking parts) provides practical learning points.

5. Support Resources: Financial advisory support from investors like Huaxing Capital suggests sellers can seek similar resources to navigate market trends.

Star Atlas's technology offers insights into production needs and commercial opportunities, particularly in smart manufacturing.

1. Production Needs: Applications in smart manufacturing and warehousing/logistics demonstrate tasks like moving boxes and picking parts, inspiring factories to optimize production line design.

2. Commercial Opportunities: The company's exploration of upstream/downstream co-creation allows factories to participate in collaborations, such as supporting product design or digitalization initiatives.

3. Digitalization Insights: The Real2Sim2Real spatial intelligence engine's ability to reconstruct scene geometry and address data issues illustrates efficient large-scale data management for factories.

4. Technical Highlights: The EFM-1 model's generalization capability suits customized production; the imitation learning algorithm reduces data requirements, boosting efficiency.

5. E-commerce Advancement: Application in cross-border e-commerce highlights the e-commerce trend, suggesting factories expand sales channels accordingly.

Star Atlas's industry trends and new technologies offer service providers solutions for client pain points.

1. Industry Trend: Embodied intelligence is a key focus in AI's physical applications, with a vast market potential; investors like Ant Group project a multi-trillion-yuan blue ocean market over the next decade.

2. New Technology: The EFM-1 embodied foundation model and Real2Sim2Real engine break new ground in solving data scarcity, enabling service providers to integrate or develop similar solutions.

3. Client Pain Points: The Real2Sim2Real engine's ability to expand one data point into thousands directly addresses data shortage issues, offering efficient solutions.

4. Innovation Direction: The imitation learning algorithm's low data requirement can inspire service providers to optimize processes; the R1 hardware's full-scene capability demonstrates application potential.

5. Development Prospects: Commercialization in areas like warehousing/logistics suggests service providers monitor trends to offer customized services.

Star Atlas's commercial needs and platform strategies provide marketplace operators with references for merchant acquisition, operations, and risk mitigation.

1. Platform Demand: Applications in cross-border e-commerce and warehousing/logistics indicate demand for platform services, suggesting opportunities for merchant acquisition and partnerships.

2. Platform Strategy: The company's rapid implementation through upstream/downstream co-creation and clear deployment paths offers lessons in operational management for platform operators.

3. Merchant Acquisition Opportunity: The funding event's attraction of multiple institutions can help platforms attract similar enterprises; the R1's tasks (e.g., door operation) provide collaboration points.

4. Risk Mitigation: The Real2Sim2Real engine's solution to data scarcity reduces platform operational risks; investor views on early-stage data dependency warrant attention.

5. Efficiency Improvement: The imitation learning algorithm's high success rate inspires platforms to optimize processes and enhance service speed.

Star Atlas's advancements in embodied intelligence reveal new industry dynamics, problem-solving approaches, and business model insights for researchers.

1. Industry Dynamics: Development of the EFM-1 embodied foundation model and the world's first Real2Sim2Real spatial intelligence engine breaks industry standards, indicating technological evolution.

2. New Problems: The Real2Sim2Real engine addresses data scarcity, offering research insights; the imitation learning algorithm's minimal data requirement (40 demonstrations) challenges traditional methods.

3. Policy Implications: Investors like Hillhouse Capital emphasize the importance of data accumulation and algorithm iteration, hinting at policy support directions; significant social value necessitates regulatory guidance.

4. Business Model: Exploration of applications in smart manufacturing, home use, and other scenarios, coupled with rapid commercialization through co-creation, provides a case study for researchers to analyze pathways.

5. Innovation Details: The DP3 3D diffusion strategy's strong generalization in real-world tasks and the R1 hardware's all-terrain capability offer valuable research case studies.