NVIDIA, Alphabet, and Google have announced a series of joint initiatives at GTC aimed at accelerating the development of agent-based AI, physical AI, and transforming key sectors such as healthcare, energy, and industrial manufacturing. This collaboration reinforces a strategic relationship that dates back to the early days of Android and now extends into the future of advanced Artificial Intelligence.
Advancements in Robotics, Health, and Energy
Engineering teams from Google DeepMind, Isomorphic Labs, Intrinsic, and X (Alphabet’s disruptive projects division) are working hand-in-hand with NVIDIA developers to create solutions that range from robots with advanced manipulation capabilities to platforms for drug discovery and optimizing electrical grids.
To achieve this, they are leveraging technologies such as NVIDIA Omniverse™, NVIDIA Cosmos™, and NVIDIA Isaac™, which allow for simulating and modeling complex physical environments and generating synthetic data to accelerate AI model training.
Cutting-Edge Infrastructure on Google Cloud
Google Cloud will be one of the first providers to adopt the new NVIDIA solutions announced at the event: NVIDIA GB300 NVL72 and the NVIDIA RTX PRO 6000 Blackwell Server Edition GPU. This infrastructure will enable Google Cloud to offer high-performance instances for companies looking to deploy AI factories capable of real-time training and inference scaling.
Additionally, Google Cloud has already launched its A4 and A4X virtual machines, the first based on the NVIDIA B200 and GB200 architectures, with general availability of A4 and upcoming announcements for A4X.
Responsible AI and Open Models
One of the standout announcements is that NVIDIA will be the first external user to implement SynthID, the watermark technology developed by Google DeepMind that allows for the identification of AI-generated content (images, videos, text, and audio) without affecting quality. This tool will be integrated into the NVIDIA Cosmos foundational models available at build.nvidia.com as a measure to combat misinformation and protect intellectual property.
On another front, both companies have collaborated to optimize Gemma, Google’s family of lightweight, open models to function on NVIDIA GPUs. Thanks to the NVIDIA AI platform, Gemma is offered as a highly optimized microservice through NVIDIA NIM, boosting its inference performance.
Next-Generation Robotics
The company Intrinsic, a subsidiary of Alphabet, has collaborated with NVIDIA to integrate more intuitive workflows into Intrinsic Flowstate, supporting foundational robotic manipulation models developed with NVIDIA Isaac. The goal is to expedite the development of robotic applications and facilitate the transition from digital twins to real hardware.
Moreover, NVIDIA and Google DeepMind, in collaboration with Disney Research, have announced Newton, a new open-source physics engine based on NVIDIA Warp technology and compatible with the MuJoCo simulator. Newton will enable acceleration of up to 70 times in machine learning processes in robotics compared to current solutions.
Applied Innovation in Healthcare and Energy
Isomorphic Labs, the company founded by Demis Hassabis (CEO of Google DeepMind), is utilizing Google Cloud infrastructure and NVIDIA GPUs to revolutionize drug discovery through AI.
On the other hand, Tapestry, another X project, is working with NVIDIA to develop AI-based solutions to enhance the reliability and sustainability of electrical networks, addressing challenges such as the integration of renewable energy sources and increasing capacity to meet the growing demand from data centers and AI applications.
The Infrastructure of the Future is Already Underway
Google Cloud and NVIDIA have worked together to optimize open-source frameworks such as JAX and MaxText, allowing for the efficient scaling of large models across clusters of thousands of GPUs.
Companies interested in learning more about these developments can visit Google Cloud’s booth (number 914) at GTC. This alliance promises to be key to developing the next generation of AI and robotics applications, bringing advanced solutions to more industries and users worldwide.