This $5.5 billion robotics startup built a school for humanoids

Before humanoid robots can take on their designated tasks, they must undergo rigorous training. Apptronik, a Texas-based startup with backing from industry giants like Google and Mercedes, has launched a state-of-the-art facility dubbed 'Robot Park' in Austin. Spanning nearly 90,000 square feet, this facility serves as a training ground for Apptronik's Apollo robots, where they refine their skills in activities such as loading boxes onto conveyor belts and sorting toys. During these training sessions, the robots are typically operated remotely, with human operators closely monitoring and guiding their actions. The data collected from these real-world exercises is crucial for enhancing the AI models that power the robots, aiming to make them valuable assets in various settings, including factories, service industries, and domestic environments. Jeff Cardenas, cofounder and CEO of Apptronik, describes this innovative training center as a 'robot learning playground.' He emphasizes the importance of generating quality data for the development of autonomous robots, a challenge that has historically hindered progress in the field. Unlike AI chatbots, which benefit from vast amounts of online text and images for training, robots have lacked comparable resources until now. In a parallel effort, Elon Musk has proposed a similar concept for Tesla's Optimus robots, envisioning an 'Optimus Academy' where robots can test their abilities and contribute to a data cycle for continuous improvement. Apptronik originated from the Human-Centered Robotics Laboratory at the University of Texas in 2016, with a mission to commercialize technology that emerged from the DARPA Robotics Challenge aimed at creating robots capable of functioning in disaster scenarios. Initially focused on developing robot components for other companies, Apptronik's vision has evolved to create a versatile humanoid robot. The startup has secured approximately $1 billion in funding, achieving a valuation exceeding $5.5 billion. Notably, Mercedes utilizes Apollo robots for basic factory tasks, such as assembling components, while DeepMind employs them to enhance Google's Gemini Robotics. The company introduced the first Apollo model in 2023 and is currently advancing to Apollo 2, which boasts improved batteries, motors, and sensors. Designed for data collection and preliminary customer trials, the Apollo 2 stands around six feet tall, can lift 55 pounds, and operates for four hours. Meanwhile, development is underway for Apollo 3, which will be marketed for commercial applications, though Cardenas did not disclose a launch timeline. Cardenas envisions the humanoid robotics market evolving through three phases: demonstrating functional technology, proving commercial viability, and ultimately scaling into profitability. He suggests the industry is now transitioning into the second phase. 'Humanoids represent the personal computer revolution of our era,' he remarked. 'If that's accurate, we're currently in the early 1980s, akin to the development of word processing and spreadsheet applications.' The race to develop humanoid robots is shifting from laboratory demonstrations to early-stage commercial testing. For instance, Figure AI from San Jose, valued at $39 billion, is starting to deploy its robots in logistics settings, while 1X in Palo Alto plans to deliver over 10,000 humanoids to consumers soon. Oregon's Agility Robotics is also making strides, with its Digit robot already operational in multiple facilities, including those of Amazon and Toyota. What sets Apptronik apart is its dual approach of developing both legged and wheeled versions of Apollo. Cardenas believes that legged robots hold the most promise for future capabilities, enabling them to perform any task a human can. However, wheeled robots may see quicker deployment due to their enhanced safety features. Cardenas has ambitious plans for expanding Robot Parks globally, aiming to make these training centers accessible to the public for a glimpse into the future of robotics.