NVIDIA and Alphabet Team Up to Shape the Future of
NVIDIA and Google’s AI Initiatives
New Collaborations to Enhance AI
On March 18, 2025, NVIDIA, Alphabet, and Google strengthened their longstanding partnership by announcing new initiatives focused on advancing artificial intelligence (AI). These initiatives aim to improve access to AI tools, expedite the development of physical AI, and transform key industries like healthcare, manufacturing, and energy.
Alphabet’s engineers and researchers will collaborate closely with NVIDIA’s technical teams to use AI for various groundbreaking projects. This includes creating robots with improved grasping abilities, revolutionizing drug discovery, and optimizing energy grid operations. The collaboration will utilize innovative platforms such as NVIDIA Omniverse™, NVIDIA Cosmos™, and NVIDIA Isaac™, with significant discussions expected at the NVIDIA GTC global AI conference.
Adoption of Cutting-Edge NVIDIA Technology
To support its AI research and production, Google Cloud will be one of the first to use NVIDIA’s latest GB300 NVL72 rack-scale solution and the powerful RTX PRO™ 6000 Blackwell Server Edition GPU, both unveiled during the GTC event. This partnership boosts Google Cloud’s AI capabilities and enhances performance for its customers.
NVIDIA will also be the first external entity to adopt SynthID, Google DeepMind’s AI watermarking technology designed to protect intellectual property by identifying content generated by AI. This technology ensures greater content transparency, preserving the integrity of outputs from NVIDIA Cosmos models.
Fostering Responsible AI Development
As part of their collaboration, Google DeepMind and NVIDIA are focused on building trust in generative AI by ensuring content transparency. They plan to use SynthID to embed digital watermarks in AI-generated media—such as images, audio, text, and video—to combat misinformation.
Additionally, NVIDIA is working with Google DeepMind to optimize Gemma, a series of lightweight, open models developed for AI applications. This partnership enhances the accessibility of Gemma for developers by utilizing the NVIDIA AI platform.
Innovations in Robotics
Intrinsic, a company under Alphabet, aims to make AI-powered robotics more adaptive and valuable across various industries. Currently, most industrial robots require manual programming, which is complex and costly. Collaborating with NVIDIA, Intrinsic has developed more intuitive workflows that reduce application development time and enhance the responsiveness of robots.
Moreover, at the GTC conference, Intrinsic will introduce an early streaming connection between its Flowstate platform and NVIDIA Omniverse. This advancement enables real-time visualization of robotic systems across different platforms, paving the way for more efficient and adaptable robotics.
Addressing Real-World Challenges with AI
Isomorphic Labs, a venture by Google DeepMind, is leveraging AI for drug discovery. They have created a cutting-edge drug design engine powered by NVIDIA GPUs on Google Cloud. This infrastructure provides the scale and performance necessary to produce innovative AI models for advancing human health.
Furthermore, X, the moonshot division of Alphabet, is collaborating with NVIDIA on Tapestry, a project aimed at creating AI-powered solutions for a more sustainable electric grid. Key challenges, such as integrating new energy sources and expanding grid capacity, will be addressed through this joint effort.
Advancements in AI Infrastructure
To maintain leadership in AI infrastructure, Google Cloud will introduce new instances featuring NVIDIA’s Blackwell GPUs. The new models will significantly boost AI performance, enabling advancements in various sectors such as healthcare, manufacturing, and retail.
Google Cloud previously became the first cloud provider to offer NVIDIA B200 and GB200-based instances, allowing customers to harness the powerful capabilities of Blackwell architecture. The availability of the A4 virtual machines will enable users to fully utilize the advantages of the NVIDIA AI Hypercomputer.
Streamlining Open-Source Frameworks
NVIDIA and Google have also collaborated to enhance popular open-source frameworks like JAX and MaxText for scalable machine learning. By co-developing optimizations, they aim to improve the efficiency and performance of AI training across large GPU clusters.
Attendees of the GTC conference can find more information about this exciting partnership at the Google Cloud booth, where they will learn about the groundbreaking work initiated by Alphabet and NVIDIA in the field of AI.
