Group 1 - RoboScience completed nearly 200 million yuan angel round financing led by JD, with participation from other investors, and plans to showcase its strategy at the upcoming World Robot Conference [1] - Zhijidynamics launched a full-size humanoid robot LimX Oli, starting at 158,000 yuan, featuring 31 degrees of freedom and modular design [1] - Giant Legend responded to its collaboration with Yushu Technology, focusing on consumer-grade robots for home use and entertainment, with a funding plan to expand retail channels [3] Group 2 - LG Energy Solution signed a $4.3 billion battery supply agreement with Tesla for lithium iron phosphate batteries, with details on the application yet to be disclosed [4]

Group 1 - Shenzhen's Huaqiangbei market has been instructed to prohibit the sale of power banks without 3C certification, leading to merchants actively removing non-compliant products from their inventory [1] - Xiaomi is urgently recruiting for its second factory to meet the production demands for the YU7 model, with wait times for the standard version reaching up to 15 months [2] - Intel is considering a strategic shift in its wafer foundry business, potentially skipping the 18A manufacturing process for new clients and directly adopting the 14A process, which may result in significant asset write-downs [3] Group 2 - The CMG World Robot Skills Competition will feature a live event showcasing a robot dog, "Black Panther 2.0," competing in unprecedented challenges, including a human-robot sprint [4]

Group 1 - The Shanghai Airport has clarified that from June 28, travelers are prohibited from carrying power banks without 3C certification, but this restriction does not apply to camera batteries, which will follow previous regulations [1] - Lexus has officially commenced construction of its new energy project in Jinshan District, Shanghai, with an expected completion date in August 2026 and an annual production capacity of 100,000 units [1] - The Dutch government and Groningen province plan to invest €200 million to build an artificial intelligence factory, which is expected to be operational by 2026 [1] Group 2 - Li Auto has announced a new organizational structure, merging its R&D and sales groups into a new "Smart Vehicle Group," with President Ma Donghui leading the group and reporting to CEO Li Xiang, who will focus more on AI initiatives [1]

Group 1 - The world's largest diameter underwater shield tunnel has successfully completed the Yellow River section excavation, marking a significant milestone in the construction of this project [1] - The project team has implemented innovative techniques to adapt to the unique geological conditions of the Yellow River basin, including increasing the shield machine's cutter head opening rate and installing a high-flow central flushing system [1] - The "Shanhe" shield machine has completed 1,376 rings and 2,752 meters of excavation, achieving 83% of the construction progress [1] Group 2 - Longxin Zhongke has partnered with Beihang University to launch China's first cross-satellite instant data service system based on Longxin, aimed at providing reliable navigation and communication coverage in complex environments [1] - This system is expected to support various sectors, including low-altitude economy, smart oceans, and emergency management, extending its influence to countries along the Belt and Road Initiative [1] Group 3 - Chuangye Huikang Technology Co., Ltd. has signed a cooperation agreement with Zhejiang University to establish a joint laboratory focused on AI applications in medicine, particularly centered around medical large models [1] - The laboratory will be located at Chuangye Huikang's headquarters and will concentrate on research and transformation of AI+medical innovations [1] Group 4 - Huawei is set to launch the new generation commercial tablet, the Qingyun C5, which features an 11.5-inch 120Hz full-screen display and runs on the HarmonyOS operating system [1]

Group 1 - The UK has developed a new type of robotic skin made from soft and low-cost gel material, capable of sensing pressure and temperature, and distinguishing multiple contact points, which can be applied in various industries including humanoid robots and disaster rescue [1] - GAC Group's smart factory in Indonesia has officially commenced production, starting with a capacity of 20,000 units and plans to gradually expand to 50,000 units, covering both pure electric and hybrid platforms [1] - Meta has launched a new smart glasses model, Oakley Meta HSTN, starting at a price of $399, designed to enhance visibility for athletes under changing light and weather conditions [1] - Lantu's FREE+ model has rolled off the production line and will begin pre-sales on June 24, with a reported cumulative sales of 46,075 units from January to May, marking an 85% year-on-year increase [1]

Group 1 - Gree Electric has achieved a significant milestone with approximately 30% of its air conditioning products utilizing self-developed chips, which are also applied in commercial air conditioning, smart equipment, and industrial robots [1] - TSMC's construction plans for new factories in Japan and Germany may face adjustments due to supply chain issues and underperformance in local market demand, particularly in the automotive sector [1] - Shanghai Hanke Technology has signed a contract to establish a high-purity process system R&D and smart manufacturing base in Songjiang, focusing on advanced technology development and intelligent production [1] Group 2 - The global wafer foundry industry is experiencing a seasonal revenue decline of about 5.4% in Q1 2025, attributed to preemptive inventory buildup and the continuation of China's old-for-new subsidy policy, resulting in a total revenue of approximately $36.4 billion [1]

扫码订阅 AIGC 产业日报， 3.【Meta Token-Shuffle登场：自回归模型突破瓶颈，可AI生成 2048×2048 分辨率图像】报道称Meta AI创 新推出Token-Shuffle，目标解决自回归（Autoregressive，AR）模型在生成高分辨率图像方面的扩展难 题。在语言生成方面，自回归模型大放异彩，近年来也被广泛探索用于图像合成，然而在面对高分辨率 图像时，AR模型遭遇瓶颈。不同于文本生成仅需少量token，图像合成中高分辨率图片往往需要数千个 token，计算成本随之暴增。这让许多基于 AR 的多模态模型只能处理低中分辨率图像，限制了其在精细 图像生成中的应用。尽管扩散模型（Diffusion Models）在高分辨率上表现强劲，但其复杂的采样过程和 较慢的推理速度也存在局限。（搜狐） 4.【Adobe发布Firefly Image Model 4模型：AI生图再升级】Adobe发布博文，推出Firefly Image Model 4和 Firefly Image Model 4 Ultra两款文本生成图像AI模型，并预告针对Photoshop和Illustrator的Crea ...