Advancing Cloud With AI: A Recap of Google Cloud Next 2023
Having tasted profits for the first time in H1 2023, Google Cloud is giving it all for a cloud-based push for AI.
- Google Cloud, which earned a record $8 billion in Q2 2023, seeks to close the gap with Amazon Web Services and Microsoft Azure.
- At Google Cloud Next 2023, generative AI took center stage with the launch of a host of tools and infrastructure.
- Google Cloud also formed a crucial partnership with NVIDIA for its GPUs and other AI tools and services.
The first in-person Google Cloud Next conference since 2019 concluded yesterday. The event marked the launch of more than a dozen new innovative AI infrastructure tools to its vast portfolio, among other development and analytics tools.
Clearly, Google Cloud, which earned a record $8 billion in Q2 2023, seeks to close the gap with Amazon Web Services and Microsoft Azure, two of the biggest cloud vendors globally. Generative AI was also a central theme at the Google I/O 2023 conference, where the company introduced PaLM 2, updated Bard, and announced AI-integrated search, Android Studio, Play Store, Workspace, Maps, and more.
Google is now replicating what it did with consumer-centric products and services at Google I/O with enterprise customers. And why not? Google’s partners are equally up for it. “Walking around the floor now, I go to every booth, and there’s at least four or five examples on every booth,” said Bronwyn Hastings, corporate vice president, global ISV partnerships and channels at Google.
“And every time I go around, they’re demoing to me the examples they’ve already got now with generative AI. So there’s a lot of excitement because customers are asking them. Customers are asking us. So they’re preparing themselves for that answer.”
Google CEO Sundar Pichai highlighted how crucial AI will become over the next decade. “As a company, we’ve been preparing for this moment for some time. And for the last seven years, we’ve taken an AI-first approach, applying AI to make our products radically more helpful, Pichai said.
Google’s focus on AI and attempts to ensure it isn’t caught off guard as it was late last year after the launch of ChatGPT is also evident from Google Cloud CEO Thomas Kurian talking more about AI and cloud itself during his 90-minute keynote address.
Well, Google also has an answer for that — its dedicated hardware and infrastructure backbone integrated with a custom software stack. For instance, Kurian pointed out that Google 30 can offer 30 different types of accelerators in the Google Cloud Platform.
Let’s look at what else Google has in store at Google Cloud Next 2023.
Google Cloud Next Key Takeaways
1. Vertex AI
Vertex AI, Google’s unified machine learning (ML) platform, got a major boost. The foundational ML model has been generally available for over two years since May 2021, with a limited number of ML models. The company is now expanding its scope to over 100 models, including open-source, first-party and third-party ones.
Some newer additions include Meta’s Code Llama and Llama 2, Technology Innovation Institute’s Falcon LLM, and Anthropic’s Claude 2. Google Cloud also updated internally developed models, such as PaLM, to enable higher quality outputs, a 32,000-token context window for analysis of large documents. Kurian claimed Codey and Imagen now deliver better performance and higher quality images, respectively.
Vertex AI also features multi-turn search, which allows users to ask follow-up questions, summarization functionality, and a Grounding feature.
With the Grounding feature in vertex AI, Google is essentially trying to avoid the backlash from AI hallucinations, or in other words, any mistake that generative AI tools end up making. It would also serve as a way to instill confidence in the generated results.
See More: Bridging the Gap Between IT and IT Decisions: A Look at SpiceWorld 2023 Sessions
2. Duet AI
Following the introduction of Duet AI in Workspace, the search giant is now introducing Duet AI in Google Cloud. Duet AI will function as an “expert coder, a software reliability engineer, a database pro, an expert data analyst, and a cybersecurity adviser.”
Duet AI in Google Cloud, seen as an answer to Microsoft 365 Copilot, is currently in preview and will be generally available later in 2023.
Besides Google Cloud, Duet AI is also coming to BigQuery. Duet AI in BigQuery is designed to free up users to focus on the logical outcome by generating full functions and code blocks. It can also suggest and write Python code and SQL queries.
VP of Data Engineering, Aritzia, said, “Duet AI in BigQuery provides contextual awareness and extends our investment in Google Cloud’s integrated data platform. We see this as an architectural advantage, eliminating the need to train, host, and manage custom models.”
3. Infrastructure announcements
-
Cross-Cloud Network
Google introduced the Cross-Cloud Network to enable seamless and secure connectivity between clouds and on-prem locations. The networking platform would allow organizations to build distributed applications and deliver internet-facing ones.
Mark Lohmeyer, Google Cloud’s VP and GM for compute and ML infrastructure, “It’s really focused on helping ease the operational requirements of running in these multi-cloud environments and enable them to focus on running their business and supporting their applications, as opposed to just operating the network.”
Google said Cross-Cloud Network can help organizations reduce network latency by 35% and total cost of ownership by 40%.
-
Google Distributed Cloud
Google updated its Distributed Cloud offering with new AI and data enhancements. Through the integration with Vertex AI, teams would be able to run AI and data analytics workloads at the edge.
-
Cloud tensor processing unit (TPU) v5e
Google calls TPU v5e its “most cost-efficient, versatile, and scalable” purpose-built AI accelerator. The company claims it can deliver a 2x improvement in training performance per dollar and up to a 2.5x improvement in inference performance per dollar over its predecessor, TPU v4.
-
A3 virtual machines
Google’s AI cloud infrastructure now incorporates NVIDIA GPUs. The two companies are joining forces for A3 VMs powered by NVIDIA’s H100 Tensor Core GPUs. The supercomputing offering is expected to deliver 3x faster training and 10x greater networking bandwidth than the previous generation.
4. Google Cloud-NVIDIA partnership
American semiconductor major NVIDIA is among the few companies that have been able to capitalize on the enormous potential of AI-driven development. The company’s technological readiness, clubbed with the right deals, has propelled its share price and market capitalization to rise over 244% in 2023.
NVIDIA CEO Jensen Huang shared the stage with Kurian at Google Cloud Next as part of its partnership with Google Cloud. “Generative AI is revolutionizing every layer of the computing stack. Our two companies with two of the most talented computing science teams in the world to re-invent cloud infrastructure for generative AI,” Huang said.
“We’re starting at every layer. This is a reimagining of the entire stack.”
Besides the NVIDIA H100 GPUs for A3 VMs, NVIDIA is also powering the Vertex AI platform with H100 GPUs. Under the expanded partnership, Google will get access to the NVIDIA DGX GH200 AI supercomputer. At the same time, its customers can leverage the NVIDIA DGX Cloud suite of AI supercomputing tools and software through a web browser and NVIDIA AI Enterprise on Google Cloud Marketplace.
“We are in an entirely new era of digital transformation, fueled by gen AI,” Kurian said. “This technology is already improving how businesses operate and how humans interact with one another. It is changing the way doctors care for patients, the way people communicate, and even the way workers are kept safe on the job. And this is just the beginning.”
Is Google well poised to take over Azure? Share your thoughts on LinkedIn, X (Twitter), or Facebook. We’d love to hear from you!
Image source: Shutterstock
MORE ON EVENTS
