Humanoids are steadily coming to life, moving beyond static frames and semi or fully controlled machines guided by humans. Apptronik, one of the leaders in AI-powered humanoid robotics, recently announced an exciting new partnership with Google DeepMind’s robotics team to create truly intelligent and autonomous robots.
The collaboration aims to advance the development of humanoid robots designed to operate in dynamic environments and assist humans in various tasks.
Jeffrey Cardenas, CEO and co-founder of Apptronik, said, “By combining Apptronik’s robotics platform with Google DeepMind robotics team’s unparalleled AI expertise, we’re creating intelligent, versatile and safe robots that will transform industries and improve lives.”
Robotics company Apptronik, known for their Apollo and Astra robots, have joined forces directly with Google DeepMind.
DeepMind is, of course, the AI giant behind AlphaFold that saved an estimated billion years of research time in protein folding, GenCast that can accurately… https://t.co/8PvrvoZSOI pic.twitter.com/HsnATkTxqW— The AI Veteran (@TheAIVeteran) December 19, 2024
Withthis collaboration, the duo looks to build robots that can perform a wide array of tasks in environments designed for humans, enhancing productivity and safety in various sectors.
Keerthana Gopalakrishnan, a researcher at Google DeepMind, commented on this and said, “Very few embodiments offer a frontier to push AI towards human-level capability in the physical world as humanoids, and they posit an AGI-hard problem in robotics!”
Apptronik, founded in 2016 out of the Human Centered Robotics Lab at the University of Texas, Austin, has developed 15 robotic prototypes, including NASA’s Valkyrie robot. Its flagship humanoid robot, Apollo, is designed for physically demanding tasks in industrial settings. At 5 feet 8 inches tall and 160 pounds, Apollo balances functionality and safety for human collaboration.
Google DeepMind’s robotics team focuses on AI-powered systems capable of reasoning and real-world action. Their work revolves around advancements in machine learning, engineering, and robotics, utilising state-of-the-art AI models like Gemini.
The partnership follows Apptronik’s recent collaborations with GXO and Mercedes-Benz, among others, as it continues to expand its industry footprint.
Meanwhile, Ram Ahluwalia, CEO and founder of Lumida, speculated on X whether this partnership could lead to Google acquiring Figure.
Apptronix’s ties with NVIDIA in March of this year further escalated this collaboration. Apollo’s main computing system, including onboard NVIDIA Jetson AGX Orin and Jetson Orin NX modules, took off from the partnership.
This enabled the AI-powered robot to efficiently leverage cutting-edge models such as GR00T. The bot could perform a wide variety of tasks, while its humanoid form allowed it to learn from human demonstrations.
Google Gemini Finally Gets a Body
“2025 will be a very interesting year, and Google wants to own it,” said Linas Beliūnas, the director for revenue at Zero Hash.
Google DeepMind recently integrated Gemini 2.0 Flash Thinking, an advanced reasoning model with capabilities that significantly enhance robotics applications. Its spatial reasoning abilities are being actively tested in robotics, enabling improved navigation and task execution in human-centric environments.
It also supports multimodal inputs like images, video, and audio, as well as outputs such as text, generated images, and multilingual speech, making the model versatile and ideal for dynamic, real-world robotic tasks.
The integration of ChatGPT into robotics highlights another major milestone in combining conversational AI with physical robots. Apptronik’s partnership with Google DeepMind could yield similar breakthroughs, enabling robots to reason and interact with their environments intelligently.
With the high rise in conversations following AGI, there have definitely been many questions about whether robotics is the path to AGI and whether Google Deepmind could be ahead in the game—a race every big tech company is invested in.
While Optimus has made strides, the collaborative expertise of Apptronik and Google DeepMind may set a new benchmark, fostering intelligent, versatile robots that transform industries.
The Rise of the Planet of the Humanoids
This collaboration comes amidst notable milestones in robotics this year. Companies like Meta and NVIDIA have shifted focus to developing advanced robotic systems.
Additionally, China’s Agibot has started mass producing humanoids as Figure AI ships its first AI robots. Interestingly, in November of last year, the Ministry of Industry and Information Technology (MIIT), which oversees the industrial sector of China, stated that in 2025 the country will achieve mass production of humanoid robots. Well, that prediction seems to be holding some water now.
Meta, for instance, has pivoted from the metaverse to robotics, exploring innovations in touch perception and human-robot interaction. These efforts aim to enhance dexterity and decision-making capabilities in robots.
NVIDIA has introduced HOVER, a neural network with 1.5 million parameters, optimised for humanoid robot locomotion and manipulation. Earlier this year, NVIDIA chief Jensen Huang said that in the near future, humanoid robots will be available for $10-20,000.
This year also saw major advancements in Tesla’s Optimus, which has contributed to significant growth in robotics investments and deployments.
In retrospect, OpenAI’s investment in Figure, alongside major players like Microsoft, NVIDIA, Bezos Expeditions, and others, reinforces the company’s interest in automating blue-collar jobs. This development posed a direct threat to Musk’s Optimus.
“We’ve always planned to come back to robotics, and we see a path with Figure to explore what humanoid robots can achieve when powered by highly capable multimodal models,” said Peter Welinder, VP of product and partnerships at OpenAI.
In addition, Brett Adcock, the founder of FigureAI robots, also hinted at the universal adoption of robots as an essential commodity in the future.
“Every human will own a robot in the future, similar to owning a car/phone today,” said Adcock.
At the beginning of this year, Google DeepMind released three robotics research systems—AutoRT, SARA-RT and RT-Trajectory—that aided robots in making faster decisions and better understand and navigate their environments. The models also helped with data collection, speed, and generalisation.
Furthermore, insights from UBTECH’s Walker S1 indicate that competition is also intensifying in areas such as balance, precision, and real-world task execution. While Tesla’s Optimus has been the talk of the internet from time to time, UBTECH’s Walker S1’s understated design and less dramatic capabilities contribute to its lower visibility in popular media.
Recently, the bot has been integrated into EV maker NIO’s advanced manufacturing factory base F2 and has been successfully performing several quality and inspection tests on a production level.
The post Apptronik and DeepMind Give Humanoids a Brain While Others Stay in the Frame appeared first on Analytics India Magazine.
