Ambi Robotics

Introducing AmbiStack. Stack to anything. AmbiStack is a multi-purpose robotic system that automates the complex process of stacking items onto pallets or into containers, optimizing space with maximum utilization. It's the smarter way to stack for warehouse operations looking to significantly improve efficiency, reliability, and worker safety while reducing transportation costs by autonomously stacking and packing items densely. The system is powered by breakthrough Sim2real Reinforcement Learning to reason about the most efficient stacking of unfamiliar items, adapting to real-world conditions on-the-fly to maximize pallet and container utilization. Get the details in our press release: https://lnkd.in/geUKW4pH Explore the new AmbiStack solution: https://lnkd.in/g85xRAY2 #HandleMore

It's been an exciting week at #PROMAT2025! And we're officially halfway through the show. Come visit our team at booth N8125. We invite you to see AmbiSort live and meet with our solutions experts to learn how AI-powered robotic sorting and stacking solutions empower your warehouse teams to #HandleMore.

4 more days until #PROMAT2025! See AmbiSort A-Series live, now powered by PRIME-1. Secure your demo with our solution experts at booth #N8125: https://lnkd.in/gDDWEGmY #PowerUp #HandleMore

More speed. More precision. More power. AmbiSort A-Series, now powered by PRIME-1. See it live at #PROMAT2025. Schedule a demo with our solution experts. Booth #N8125: https://lnkd.in/eSMxAswY #HandleMore

More power with PRIME-1. AmbiSort A-Series, now powered by PRIME-1—our most advanced AI yet. More speed. More intelligence. More power than ever before. See it. Believe it. Live at #PROMAT2025. Schedule your demo today: https://lnkd.in/gDDWEGmY #HandleMore

The real robot revolution needs real data. Today, there’s a 100,000x "data gap" between general vision-language models and robotics. AI models trained on internet-scale data have transformed industries, but robots need systematic, real-world experience to master physical tasks. At NVIDIA GTC, Ken Goldberg—Co-Founder and Chief Scientist of Ambi Robotics—will break down why data, not just computation, is the key to unlocking industrial robotics at scale. Learn how Ambi Robotics is contributing to this effort. With 200,000+ hours of real robot data (20x more than recent AI unicorns), our fleet of AI-powered robotic systems have sorted over 100 million consumer packages. Each sort is a step toward closing the data gap for a practical subclass of robot skills, enabling a new generation of real industrial robots. This is igniting a real robot revolution. Don't miss this talk at #GTC25. Register with NVIDIA today: https://lnkd.in/g5cfRP3F

This week during AI After Hours at GitHub HQ, our Foundation Model Lead, Vishal Satish, shared how Ambi Robotics is leveraging 200K+ hours of high-fidelity production data to train PRIME-1—a domain-expert foundation model designed for industrial reliability. PRIME-1 demonstrates the precision required for stereo vision and robotic pick planning while exhibiting neural scaling laws traditionally found in LLMs—validating our data-flywheel approach to smarter automation. A huge thank you to Encord, the AI After Hours team, and everyone who joined this prime conversation.

Power up for PROMAT 2025! AmbiSort A-Series, now with PRIME-1. Faster. Smarter. More powerful. See it live. Booth #N8125. March 17-20. Book your meeting today: https://lnkd.in/gDDWEGmY #HandleMore | #PROMAT2025

The next frontier of AI is real-world robotics. But achieving industrial-grade reliability demands more than what we expect from AI chatbots. 80+ robots deployed. 200,000+ operational hours. 1 BILLION+ images. All collected from real production environments handling customer parcels. We've built a critical mass of high-quality data to create highly reliable, data-driven foundation AI models. Our PRIME-1 foundation model is already showing a 19% improvement in stereo vision accuracy, and we've only used a fraction of our data. The scaling laws suggest we're just getting started. Read our technical deep-dive on PRIME-1: https://lnkd.in/gqU9jT5W

Love at first stack. 📦 💜