Manycore Tech

Introducing Aholo Open Platform’s first user: Zein from Syria! 🌍 Using Aholo, he recreated the thousand-year-old Shisi Temple in Zhejiang. As a fan of Black Myth: Wukong, Zein hopes to share traditional Chinese culture in this creative way with family and friends back home. This beautiful moment, preserved through 3D technology. 💫 youtube.com/watch?v=kCEsyA… #Aholo #3DModeling #CulturalHeritage #BlackMythWukong #DigitalPreservation

Today marks the #SpringEquinox — #Chunfen, the 4th solar term in the Chinese lunar calendar. 🌱 Longer days, warmer weather, and nature picking up speed. Perfect time to enjoy flowers, spot baby animals, and take in those spring scents. @LuxRealAI Let’s step into the season. 🌸 #LuxReal #spatialintelligence

OMG did I just see Mona Lisa clock in for a Zoom meeting? 🎨💼 When iconic paintings become 9-to-5 office workers, you can literally feel the "corporate fatigue" radiating through the canvas... The side effects of modern work life are so real, even masterpieces can't escape! #LuxReal #AIVideo #OfficeLife #Spatialintelligence youtube.com/shorts/d32RuMm…

OMG did I just see Mona Lisa clock in for a Zoom meeting? 🎨💼 When iconic paintings become 9-to-5 office workers, you can literally feel the "corporate fatigue" radiating through the canvas... The side effects of modern work life are so real, even masterpieces can't escape! #LuxReal #AIVideo #OfficeLife #Spatialintelligence

A Moment Becomes Eternity. Five hundred years old. Still blooming. 🌿🌸 Faxi Temple's ancient magnolia is one of Hangzhou's most breathtaking spring secrets — and we've captured every petal in 3D. This is history you can see, feel, and now, keep forever. #Aholo #Spatialintelligence #3DGS #3D youtube.com/watch?v=4x6KzQ…

@InZhejiang Five hundred years old. Still blooming. 🌿🌸 Faxi Temple's ancient magnolia is one of Hangzhou's most breathtaking spring secrets — and we've captured every petal in 3D. This is history you can see, feel, and now, keep forever. #Aholo #Spatialintelligence #3DGS

Zhejiang-based Manycore Tech recreated a 500-year-old magnolia flower using AI 3D scanning technology, bringing nature into the digital world. 🌸

Today, together with Zhejiang University and Unitree, we took another small step forward in the evolution of robotics: Two of our joint research papers have been accepted to CVPR 2026! 🚀 For those who don’t know, CVPR is widely regarded as the "Oscars of computer vision" — bringing together the world’s most cutting-edge AI research every single year. These two papers come from the same collaborative initiative — a major provincial sci-tech project led by Manycore Tech Inc., in partnership with Zhejiang University, Unitree, and Zhejiang Humanoid Robotics Innovation Center. The project’s mission? Exploring how we can help robots truly adapt to and understand the real world. Along the way, we’ve focused on two key research frontiers: 🤖 How robots can learn continuously 🧠 How AI can truly make sense of space #CVPR2026 #AIResearch #Robotics #SpatialIntelligence #HumanoidRobots #TechInnovation #Manycore #ZhejiangUniversity #Unitree

Here’s a clip from a martial arts short drama set in the ancient Chinese temple, Shisi Temple. Using 3D Gaussian Splatting technology, the Aholo platform by Spatial Intelligence has digitally transformed this millennium-old temple into a stunning 3D version. The result is breathtaking—stay tuned for more incredible content! youtube.com/watch?v=jgN6td…

Happy Chinese New Year！ In 2026, join ManyCoreTech as we embrace every moment—full speed ahead! 🚀✨ #ManyCoreTech #2026 #luxreal

LuxReal’s new AI Short Drama feature reimagines lost cultural relics returning home for the New Year. This year, don’t just watch the story — direct it. Be your own creator and let your imagination run wild. 🪄 Create your own AI short film with LuxReal. #LuxReal #AI #AIVideo #AIShortFilm #CreativeAI #DigitalStorytelling

Big news 🎉 Our work “Towards Physically Executable 3D Gaussian for Embodied Navigation” is accepted at #ICLR2026! We turn 3D Gaussian Splatting into a physically executable, semantically aligned world for embodied navigation: 🏠 1K annotated scenes (InteriorGS) 🧭 2M trajectory–instruction pairs (SAGE-Bench) 🚀 Fully open-sourced on Hugging Face. Website: sage-3d.github.io Data: huggingface.co/datasets/spati… huggingface.co/datasets/spati… huggingface.co/datasets/spati… Code: github.com/Galery23/SAGE-… #ICLR2026 #EmbodiedAI #3DGaussianSplatting #VisionLanguageNavigation

Big news 🎉 Our work “Towards Physically Executable 3D Gaussian for Embodied Navigation” is accepted at #ICLR2026! We turn 3D Gaussian Splatting into a physically executable, semantically aligned world for embodied navigation: 🏠 1K annotated scenes (InteriorGS) 🧭 2M trajectory–instruction pairs (SAGE-Bench) 🚀 Fully open-sourced on Hugging Face🤗 Website: sage-3d.github.io Data: huggingface.co/datasets/spati… huggingface.co/datasets/spati… huggingface.co/datasets/spati… Code: github.com/Galery23/SAGE-… #ICLR2026 #EmbodiedAI #3DGaussianSplatting #VisionLanguageNavigation

How can timeless crafts meet the future? Aholo, Manycore's open spatial intelligence platform, acts as their digital guardian. We build precise digital archives so cultural treasures aren't just remembered—they remain accessible, forever. #3DGS #Traditionalcrafts #spatialintelligence #Aholo

How can timeless crafts meet the future? Aholo, Manycore's open spatial intelligence platform, acts as their digital guardian. We build precise digital archives so cultural treasures aren't just remembered—they remain accessible, forever. #3DGS #Traditionalcrafts #spatialintelligence #Aholo

AI video has looked impressive for years. But almost none of it survives real production. Why? Most tools chase spectacle, not consistency. Frames drift. Lighting breaks. Products mutate. Logos move. That’s fine for demos. It fails in e-commerce and ads, where accuracy is non-negotiable. This is why LuxReal, built by @ManycoreTech, is interesting. It’s not guessing pixels. It understands space. 3D-native AI. Spatial + temporal consistency. Actual control. Physically correct light and motion. Same characters across shots. 360° cameras. Complex movement. Nothing breaks. AI video is finally shifting from “fun to watch” → “usable in production.” LuxReal is early, closed beta, broader access in 2026. But this is what the next phase actually looks like. Apply here: luxreal.ai/ai-video?utm_s…

The real problem with AI video isn’t creativity. It’s spatio-temporal inconsistency. Most AI videos collapse when camera angles change. Objects shift. Shapes break. Perspective fails. Why? Traditional AI video models are trained on: • Flat images • Short clips • Pattern correlations They don’t understand: • Depth • Physical space • Object permanence That’s why most AI videos look impressive but fail in real use. This is where LuxReal changes the game. Meet LuxReal: luxreal.ai/ai-video?utm_s…

China’s Silicon Valley? CNBC interviews Manycore Tech Co-founder & Chairman Huang Xiaohuang on why Hangzhou is emerging as the country’s next innovation hotspot.@CNBCi #SiliconValley #innovationtrends #hangzhou

Manycore Tech, in collaboration with Zhejiang University, has officially open-sourced SAGE-3D — the industry’s first VLN (Vision-Language Navigation) embodied learning dataset built on 3D Gaussian representations. 🌍 Now available on Huggingface, with 10,000+ global downloads and growing. What’s inside SAGE-3D: 🏠 1,000 scenes 🧩 554K+ object instances with semantic & collision data 🧭 2M trajectory–instruction pairs for VLN tasks 🔗 Now available on Huggingface🤗： VLN Data: huggingface.co/datasets/spati…()_Data Interior 3D Gaussian (USDZ): huggingface.co/datasets/spati…()_usdz Collision Mesh: huggingface.co/datasets/spati… #OpenSource #EmbodiedAI #VLN #3DGaussian #SpatialIntelligence #Robotics #AIResearch

Manycore Tech, in collaboration with Zhejiang University, has officially open-sourced SAGE-3D — the industry’s first VLN (Vision-Language Navigation) embodied learning dataset built on 3D Gaussian representations. 🌍 Now available on Huggingface, with 10,000+ global downloads and growing. What’s inside SAGE-3D: 🏠 1,000 scenes 🧩 554K+ object instances with semantic & collision data 🧭 2M trajectory–instruction pairs for VLN tasks 🔗 Now available on Huggingface🤗 VLN Data: huggingface.co/datasets/spati…()_Data Interior 3D Gaussian (USDZ): huggingface.co/datasets/spati…()_usdz Collision Mesh: huggingface.co/datasets/spati… #OpenSource #EmbodiedAI #VLN #3DGaussian #SpatialIntelligence #Robotics #AIResearch

That's so cooool