The AI Hypercomputer features performance-optimized compute, storage, and networking built over an ultrascale data center infrastructure. It supports popular machine learning frameworks such as JAX, TensorFlow, and PyTorch, and offers a range of flexible and dynamic consumption choices. Customers like Salesforce and Lightricks are already using Google Cloud’s TPU v5p AI Hypercomputer and have reported significant improvements in training speed. Google is inviting customers to request access to Cloud TPU v5p and AI Hypercomputer to innovate with AI faster, more efficiently, and more cost-effectively.
Key takeaways:
- Google has announced Cloud TPU v5p, its most powerful, scalable, and flexible AI accelerator to date, designed to address the challenges of training, tuning, and inference in generative AI models.
- The company is also introducing AI Hypercomputer from Google Cloud, a supercomputer architecture that integrates performance-optimized hardware, open software, leading ML frameworks, and flexible consumption models.
- Cloud TPU v5p can train large LLM models 2.8X faster than the previous-generation TPU v4 and is 4X more scalable in terms of total available FLOPs per pod.
- AI Hypercomputer offers performance-optimized hardware, open software with extensive support for popular ML frameworks, and flexible consumption choices, all designed to enable modern AI workloads.