From Google DeepMind’s AI developments, to Google Cloud’s AI-first infrastructure, to our Different Bets’ use of AI to sort out advanced challenges in areas like healthcare, vitality and robotics — each nook of our firm is targeted on pushing the frontier of AI and bringing its advantages to everybody.
NVIDIA is a important accomplice on this mission. As a world chief in AI and accelerated computing, NVIDIA has collaborated with Google and Alphabet for years on all kinds of tasks, together with Android, superior AI analysis, next-generation {hardware} and software program optimizations, and past — all with the purpose of constructing AI extra accessible for builders.
This week on the NVIDIA GTC world AI convention, we’re doubling down on our partnership with NVIDIA with bulletins throughout Google and Alphabet:
Google Cloud’s AI infrastructure will assist remodel companies with NVIDIA’s newest GPUs
Coaching giant AI fashions requires critical computing energy, and NVIDIA continues to boost the bar on GPU efficiency. We’re excited to announce our A4 VM is now usually accessible, based mostly on NVIDIA HGX B200 GPU — and our A4X VMs based mostly on NVIDIA GB200 NVL72 will probably be accessible quickly, too. We’re dedicated to supporting the newest Blackwell GPUs, together with the just-announced NVIDIA RTX PRO 6000 Blackwell and NVIDIA GB300. This implies quicker coaching occasions, smoother deployments and the flexibility to sort out much more advanced AI challenges.
Google DeepMind and NVIDIA will make the ability of Gemini accessible to extra folks
Google DeepMind’s Gemini is the results of many years of enormous language mannequin analysis and is our most succesful AI mannequin. This week we’ll share how fast advances in Gemini are designed to help extra builders and finish customers — and the way our collaboration with NVIDIA helps to make that doable.
The Google DeepMind staff additionally launched Gemma 3 final week, the newest in our assortment of light-weight, state-of-the-art open fashions. NVIDIA performed a key function within the improvement of those open fashions, working intently with our researchers to optimize Gemma to run on its huge accelerated computing ecosystem. This implies builders can simply entry the identical expertise that powers Gemini and implement it on any NVIDIA GPU accessible to them.
Individually, NVIDIA additionally introduced at present that it has employed Google DeepMind’s SynthID watermarking device on its Cosmos video era platform. This marks our first exterior deployment of the expertise, and we’re proud to assist instill consumer belief in AI-generated content material.
Alphabet and NVIDIA will use AI to sort out the world’s most advanced points
Our collaboration with NVIDIA isn’t nearly constructing AI instruments to assist in every day life; it’s about basically enhancing the standard of life for everybody. Which means working collectively to use AI to deal with a number of the world’s most advanced challenges. For starters, we’re teaming as much as apply essentially the most superior frontier fashions to areas from vitality to drug discovery.
- Smarter vitality grids. Tapestry, X’s moonshot for the electrical grid, and NVIDIA are researching strategies to extend the pace and accuracy of electrical grid simulations.
- Improved drug discovery. Isomorphic Labs and NVIDIA are advancing the event of recent medicines utilizing AI.
- Extra succesful robots. Intrinsic and NVIDIA are making robots extra clever and succesful, with the mixing of NVIDIA’s Isaac basis fashions for extra adaptive grasp capabilities.
- Superior robotics simulation. Google DeepMind and NVIDIA are launching MuJoCo-Warp, a brand new open-source physics simulator that may speed up robotics analysis.
We’ve at all times believed that the worth of expertise lives in its capability to learn folks in every single place. That’s been foundational to our mission and it’s what our rising collaboration with NVIDIA is constructed on. By combining Google’s experience in AI analysis and infrastructure with NVIDIA’s management in accelerated computing, we’re dedicated to strengthening the muse of AI, making AI extra accessible and useful to builders and customers, and collaborating in methods that may drive innovation for years to return.
For those who’re attending GTC this week, swing by Google Cloud’s sales space #914 and take a look at one of many many Google periods to study extra about our work collectively.
