Google unveiled a range of updates and features at its Cloud Next 2024 event aimed at enhancing customer success and momentum. Let’s delve into the key announcements and features from Google Next ’24:
Google Axion Processors
Google introduced Google Axion Processors as their first custom Arm-based CPUs tailored for data centers. These processors redefine performance and energy efficiency, surpassing current-generation x86-based instances by up to 30% in performance and up to 60% in energy efficiency.
Moreover, Axion processors, fueled by Arm Neoverse V2 CPU, elevate performance for various workloads, supported by Titanium, a custom silicon microcontroller enhancing performance and capacity by managing platform operations.
This ensures compatibility with existing applications, thanks to Axion’s adherence to the standard Armv9 architecture. Soon, services like BigTable and Spanner will harness the power of Axion processors.
Collaboration with Arm’s SystemReady Virtual Environment further streamlines the deployment of Arm workloads on Google Cloud, facilitating seamless integration with various Google Cloud services and software available on the Google Cloud Marketplace.
Availability: Customers can anticipate accessing virtual machines based on Axion processors in preview within the coming months, with full availability slated for later this year.
Updates to Google Cloud include:
Google introduced custom silicon advancements like TPU v5p and Google Axion, their first custom Arm-based CPU for data centers. Gemini 1.5 Pro, enhancing long context understanding, entered public preview.
They added grounding capabilities in Vertex AI, Gemini Code Assist for developers, and expanded cybersecurity with Gemini in Threat Intelligence, along with improvements in Google Workspace.
- AI Hypercomputer: A supercomputing architecture combining hardware, open software, ML frameworks, and flexible consumption models.
- Foundation models: Including Gemini models for multi-modal information processing and advanced reasoning skills.
- Vertex AI platform: Assisting organizations in deploying custom models and connecting them with enterprise data and processes.
- Gemini for Google Cloud: Providing AI assistance for efficient work, code management, data insights, and security threat identification.
- Gemini for Workspace: Integrated into Gmail, Docs, Sheets, etc., with enterprise-grade security.
To support AI-driven transformation, Google offers an AI Hypercomputer with TPUs, GPUs, and AI software, facilitating efficient model training and serving. Notable advancements include:
- A3 Mega: Using NVIDIA H100 Tensor Core GPUs for demanding workloads.
- NVIDIA HGX B200 and GB200 NVL72: Designed for AI, data analytics, and large language model inference.
- TPU v5p: Their most powerful AI accelerator, now generally available with increased compute power.
- AI-optimized storage options: Accelerating training speed with new caching features and Hyperdisk ML for faster model load times.
- Dynamic Workload Scheduler: Introducing calendar mode and flex start for efficient resource management.
- Google also announced advancements in analytics, databases, cybersecurity, compute, networking, and their AI ecosystem.
Google is also extending AI to edge and air-gapped environments through Google Distributed Cloud (GDC), enabling AI deployment tailored to specific organizational needs, exemplified by Orange’s use case for network performance enhancement.
Also, Google introduces new features in GDC (Google Data Centers):
- NVIDIA GPUs are now available for both connected and air-gapped configurations in GDC, supporting new GPU-based instances for efficient AI model operations.
- GKE (Google Kubernetes Engine) technology used in Google Cloud is now accessible in GDC.
- Validation of various open AI models like Gemma and Llama 2 for operation in air-gapped and connected edge environments.
- AlloyDB Omni for Vector Search enables search and retrieval on GDC for sensitive data with low latency.
Sovereign Cloud offers GDC in fully air-gapped configuration with local operations, ensuring compliance with strict regulatory requirements.
Furthermore, Google announces:
- N4 and C4, two new machine series in the general-purpose VM portfolio.
- Native bare-metal machine shapes in the C3 machine family.
- General availability of Hyperdisk Advanced Storage Pools.
Vertex AI updates
- Expanded access to over 130 models, including Gemini, Claude 3, Gemma, Llama 2, and Mistral.
- Gemini 1.5 Pro offers support for up to 1 million tokens and can process audio files, providing extensive information processing capabilities.
- Claude 3 Sonnet and Claude 3 Haiku are generally available, with Claude 3 Opus coming soon.
- CodeGemma, a fine-tuned version for code generation and assistance, is now available.
- Imagen 2.0 introduces text-to-live image capability, advanced photo editing features, and digital watermarking for AI-generated images.
For developers:
- Gemini Code Assist speeds up coding tasks with more accurate suggestions and insights, now supporting private code bases.
- Gemini Cloud Assist provides AI assistance across the application lifecycle, facilitating design, security, operation, troubleshooting, and optimization.
In cybersecurity:
- Gemini in Threat Intelligence offers deep insights about threat actor behavior using natural language.
- Gemini in Security Operations introduces assisted investigation features for faster threat detection and response.
New features and updates in Google Workspace
Google Vids: A new AI-powered video creation app designed for work purposes. It simplifies the process of creating videos with features like storyboard generation, stock media integration, and voiceover assistance.
AI Meetings and Messaging Add-on: This add-on brings AI features such as automatic note-taking and translation support for meetings and messages. It’s available for USD 10 (Rs. 832 approx.) per user per month.
AI Security Add-on: Another add-on aimed at enhancing security measures, allowing automatic classification and protection of sensitive files in Google Drive. It’s also priced at $10 (Rs. 832 approx.) per user per month.
Workspace and Vertex AI Integration: Google is facilitating the integration of custom AI agents built with Vertex AI into Workspace, streamlining workflows.
Workspace Improvements: Several enhancements are coming to existing Workspace features, including voice prompting and instant polish in Gmail.
And there’s also new building blocks and notifications in Sheets, tabs and personalization in Docs, and increased member capacity and interoperability in Chat.
Availability
All these updates and features are available immediately or will be rolled out in the coming weeks. Users can also opt for a no-cost trial to experience the new offerings.