Google Announces New Axion AI Chip, Gemini Updates, and More

Google is kicking off its Cloud Next 24 conference in Las Vegas today, and the company has lots of AI-related announcements this morning. First of all, Google just put the wraps off Axion, the company’s first new ARM-based AI CPU for data centers. Moreover, Google is also expanding access to its AI models across its cloud offerings.

“We last came together just eight months ago at Next 2023, but since then, we have made well over a year’s progress innovating and transforming with our customers and partners,” said Thomas Kurian, CEO of Google Cloud. “We have introduced over a thousand product advances across Google Cloud and Workspace.”

While Google has been quick to make its Gemini AI models available to developers and organizations, the company has also been scaling up its cloud infrastructure across 40 regions. And the new Google Axion processors that have been announced today should help to make ARM-based servers in Google Cloud even more competitive.

Google Axion AI CPUs to Deliver Better ARM Performance for Data Centers

Google’s new ARM-based Axion AI CPUs will be available to Google Cloud customers later this year. The company promises “giant leaps in performance” for general-purpose cloud workloads, as well as AI training and inferencing tasks that require powerful CPUs.

“Axion processors combine Google’s silicon expertise with Arm’s highest performing CPU cores to deliver instances with up to 30% better performance than the fastest general-purpose Arm-based instances available in the cloud today, up to 50% better performance and up to 60% better energy-efficiency than comparable current-generation x86-based instances,” explained Amin Vahdat, Vice President of Machine Learning, Systems, and Cloud AI at Google.

Google has already started to deploy some of its services such as the Google Earth Engine and the YouTube Ads platform on ARM-based servers powered by its new Axion chips. This work is still ongoing, and eventually, Google Cloud customers will be able to leverage the new AI chips when using the company’s cloud-based developer services. That includes the Google Compute Engine, Google Kubernetes Engine, Dataproc, Dataflow, Cloud Batch, and more

Gemini 1.5 Pro now available in public preview in for Cloud customers and developers

Gemini Pro 1.5, the new multimodal large language model that Google made available in early testing in February is now available in public preview in Vertex AI, the company’s enterprise AI platform. Gemini Pro is Google’s first model to offer a 1 million-token context window – tokens are the building blocks used by AI models use to process information – to improve its generative AI capabilities.

Gemini 1.5 Pro is also now available in private preview in Gemini Code Assist, which developers can use to generate and text code. Thanks to the model’s larger context window, Google says that Gemini Code Assist should provide even more accurate code suggestions.

Imagen 2.0 gets new image generation capabilities

On its Vertex AI platform, Google actually provides access to more than 130 AI models, including popular third-party models such as Claude 3, Mistral, and Meta’s Llama 2. Imagen 2.0, Google’s text-to-image model is also available on Vertex AI, and it’s being upgraded today with advanced photo editing features and text-to-live image capabilities in preview.

“Today’s preview of text-to-live image capabilities makes Imagen even more powerful for enterprise workloads,” Vahdat explained today. “This allows marketing and creative teams to generate animated images such as GIFs, and more, from a text prompt. Initially, live images will be delivered at 24 frames per second (fps) with a resolution of 360×640 pixels and a duration of 4 seconds, with plans for continuous enhancements.”

New AI powered features in Google Workspace

Back in February, Google made Gemini for Google Workspace, which was previously named Duet AI for Google Workspace available for all Google Workspace customers. Similar to Microsoft’s Copilot for Microsoft 365 and Copilot 365 subscription, Gemini for Google Workspace brings generative AI features into Gmail, Google Docs, Google Sheets, and other cloud-based apps.

Starting today, Google is bringing more AI features to Gemini for Google Workspace: Gemini is coming to Google Chat in preview where it will be able to summarize conversations, answer questions, and more. Google is also increasing the member capacity of spaces to up to 500,000 people in the coming weeks, and messaging interoperability with Slack and Teams is now generally available.

With Google Vids, Google Workspace is also getting a new AI-powered video creation app. Vids will be available through Google Workspaces Labs in June, and it already looks quite promising.

“It can generate a storyboard that you can easily edit, and after choosing a style, it pieces together your first draft with suggested scenes from stock videos, images, and background music,” explained Aparna Pappu, GM and Vice President, Google Workspace. “It can also help you land your message with the right voiceover — either choosing from one of our preset voiceovers, or using your own.”

Lastly, Google announced two new commercial add-ons for Google Workspace today, both priced at $10 per user per month. The new AI Meetings and Messaging add-on will add new features including “Take notes for me” and Translate for Me” in Google Meet. The new AI Security add-on, however, will be able to utomatically classify and protect sensitive files in Google Drive.