Paulina Twarogal
Google Cloud Next ’24, held in Las Vegas from April 9th to 11th, attracted over 30,000 attendees along with 2,500 partners. The conference showcased the latest advancements in Google Cloud technologies, offering in-depth sessions, workshops, and demos. This year’s event focused heavily on the practical uses of generative AI and its potential to transform workplaces. Over 300 stories from customers and partners illustrated how AI is reshaping various industries and streamlining business processes.
Announcements from Google Cloud are always a highlight of the conference, and this year was no exception. With a record-breaking 218 announcements, Google Cloud CEO Thomas Kurian unveiled significant advancements during his opening keynote address. So, if you couldn’t make it to the event, here’s a summary of the key highlights.
Google Cloud Next 2024 highlights
Image credit: Google Cloud
Just a year ago, GenAI’s potential for business change was still a big question mark. Today, that transformation is clear. Sundar Pichai, the CEO of Google, highlighted a key reason behind Google Cloud’s success: their significant investments in Artificial Intelligence.
He explained that because Google recognized AI’s transformative potential for companies early on, they’ve invested heavily in building the infrastructure and tools businesses need. This put Google Cloud in a leading position as companies move towards using AI platforms. And the focus on AI is paying off. Pichai proudly announced that over 60% of funded AI startups and nearly 90% of the most successful ones rely on Google Cloud.
1. AI-optimized infrastructure
Image credit: Google Cloud
Arm-based chip: Axiom
Google continues its efforts to optimize AI infrastructure with their custom Arm-based central processing unit (CPU), Axiom. This chip competes with similar chips offered by other big tech companies like AWS (Graviton), Alibaba (Yitian 710), and Microsoft (Azure Maia 100 and Cobalt 100). While Google hasn’t revealed all the technical details yet, they claim Axiom leverages Arm’s Neoverse 2 architecture and delivers a significant advantage.
Axiom chips are designed for all sorts of computing tasks and boast some impressive features. Google claims they use 60% less energy than similar chips used by competitors, while also being 60% faster. Plus, they reportedly outperform existing Arm-based chips from Amazon and Microsoft by 30%.
AI hypercomputer: TPU v5p
Google also introduced its most powerful tensor processing unit yet—TPU v5p. It can scale to tens of thousands of chips. A single pod of the v5p contains 8,960 chips. That’s nearly double the chips compared to the previous version (TPU v4). The v5p also features significant enhancements in other areas: it can perform calculations twice as fast, store three times more data, and handle even larger workloads—four times bigger than before.
This focus on powerful processors isn’t the only exciting news from Google Cloud. The company is also offering a suite of tools to help businesses get the most out of its AI infrastructure. This includes special storage solutions for AI data (Hyperdisk ML), a system for managing workloads efficiently (Dynamic Workload Scheduler), and support for powerful graphics processing units (GPUs) from NVIDIA (H100).
2. Vertex AI updates
Image credit: Google Cloud
Vertex AI Agent Builder
This is a new tool that lets anyone, even without coding experience, build and deploy AI assistants. It uses powerful language models, developer tools, and Google Search. The company also equipped it with retrieval augmented generation (RAG) and vector search to deliver accurate results and reduce hallucinations. Vertex AI Agent Builder leverages Gemini’s large language models. Google showed examples of using it for tasks like managing customer inquiries in retail or handling contracts.
More models in Vertex AI
Vertex AI, already offering 130 models, expands its capabilities with the introduction of several new AI models. This includes:
- Claude 3: A powerful new family of AI models from Anthropic.
- Gemini 1.5 Pro: Google’s latest and improved large language model.
- CodeGemma: A brand new tool to help programmers write code more easily.
- Imagen 2.0: An advanced technology that allows you to create realistic images from just a text description.
- Digital Watermarking: Imagen 2.0 now supports digital watermarking powered by Google DeepMind’s SynthID.
Gemini 1.5 Pro
Gemini 1.5 Pro is now in public preview. It can now process vast amounts of information in a single stream, including up to 1 hour of video, 11 hours of audio, codebases exceeding 30,000 lines of code, and over 700,000 words of text. Gemini 1.5 Pro strongly competes with OpenAI Sora.
3. Accelerated development with Gemini Code Assist
Image credit: Google Cloud
Gemini Code Assist is an AI code-assistance solution that helps developers write code faster and more efficiently. Google’s internal tests showed a 40% increase in development times and a 55% reduction in time spent writing new code. The tests also revealed that Google’s customer, Quantiphi, saw its developer productivity grow by over 30% thanks to Gemini Code Assist.
What’s more, Google proudly shared that Gemini Code Assist supports private codebases stored anywhere. That is on-premises servers, cloud platforms like GitHub, GitLab, and Bitbucket, or even across multiple locations.
4. AI innovations for Google Workspace
Image credit: Google Cloud
With over 3 billion users, Google Workspace is the go-to suite for individuals and enterprises. Recent enhancements like AI-powered agent Gemini in Gmail, Docs, and Sheets are proving valuable. Google revealed that 70% of users leverage Gemini’s writing suggestions in Docs, and 75% use its image creation in Slides.
Google Workspace empowers employees at leading brands like Uber, Verizon, and Woolworths. At the event, Google announced new innovations for Gemini in Google Workspace including:
- Google Vids: This AI video creation app assists with scripting, editing, and production. It suggests storyboards, stock footage, and music, even offering voiceovers. Vids will be released to Workspace Labs in June.
- AI Meetings and Messaging add-on: This collaboration tool provides features like note-taking, chat summarization, and real-time translation in 69 languages.
- New AI Security add-on: Admins can automatically classify and protect sensitive data using AI and Data Loss Prevention controls. It’s available for $10 per user/month and can be added to most Workspace plans.
5. AI-driven insights: transforming data analysis
Image credit: Google Cloud
Google has upgraded their AI-ready Data Cloud with features like LLM integration and vector matching across databases, including BigQuery and AlloyDB. This empowers data teams to use advanced AI models, such as Gemini, for tasks like improving patient care, streamlining supply chains, and boosting customer engagement. These advancements will be particularly beneficial in telecommunications, retail, and financial services.
New tools for creating effective data agents:
- Gemini in BigQuery: With Gemini in BigQuery, AI assists data teams in tasks like data preparation, discovery, analysis, and governance. Moreover, the new BigQuery data canvas allows users to create and run data pipelines with a user-friendly interface. It offers a notebook-like experience with natural language and visualizations, currently available in preview.
- Gemini in Databases: This feature simplifies migrating data from legacy systems to modern cloud databases like AlloyDB. It ensures a safe and smooth transition.
- Vector indexing: New query abilities using vector indexing directly in BigQuery and AlloyDB enable AI usage right where your data is stored. This means you get quick and accurate responses in real time.
- Gemini in Looker: Integration of data agents with workflows becomes easier, allowing users to interact with business data using next-gen AI capabilities. And it is compatible with Google Workspace.
6. AI-powered security
Image credit: Google Cloud
The rise in both the frequency and complexity of cybersecurity attacks is a growing concern. However, with GenAI, there’s hope that defenders can gain an edge. Security Agents now lend a hand at every stage of the security process: prevention, detection, and response. At the conference, Google announced new AI-driven enhancements across its security portfolio. These innovations aim to deliver stronger security outcomes and empower organizations to integrate Google into their security strategies seamlessly.
Gemini in threat intelligence: Gemini leverages natural language to provide deep insights into threat actor behavior. By analyzing larger sets of potentially malicious code, Gemini sheds light on code intent by examining interactions between modules.
Gemini in security operations: This new feature simplifies investigations by translating natural language into detections, summarizing event data, suggesting actions, and guiding users through the platform via conversational chat.
Wrap up
Google Cloud Next 2024 delivered a wealth of new announcements and insights. The event showcased cutting-edge technologies and thought-provoking discussions about the future of cloud computing. Apart from that, Google demonstrated its strong commitment to AI dominance, with substantial investments and a unified branding strategy centered around Gemini. This dedication positions GenAI as a cornerstone of Google’s competitive edge in the ever-evolving tech landscape. As organizations look to the future, these innovations will play a pivotal role in shaping how they work, collaborate, and innovate.
