Google Axion is a custom Arm datacenter CPU 

At Cloud Next 2024, Google Cloud made a number of enterprise and infrastructure announcements, including Axion and Gemini. 

On the infrastructure front, Google announced “Axion” as its “first custom Arm-based CPU designed for the datacenter.” Joining Tensor on the consumer hardware side, the company touts:

• “50% better performance than comparable current-generation x86-based VMs”
• “60% better energy efficiency than comparable current-generation x86-based VMs”
• “30% better performance than the fastest general-purpose Arm-based instances available in the cloud today”

Internally, Google is already using Axion to deploy first-party services like BigTable, Spanner, BigQuery, Blobstore, Pub/Sub, Google Earth Engine, and the YouTube Ads platform in Google Cloud. Usage will continue to expand in the future, while “Axion-based instances will be available later this year.”

Meanwhile, TPU v5p is entering general availability. The AI accelerator for training and inference has “4x the compute power per pod” compared to the previous generation.

Gemini 1.5 Pro is now available in public preview in Vertex AI with two context windows: 128K tokens and 1 million tokens. New is the ability to process audio, including standalone files and videos with audio.

Customers can process vast amounts of information in a single stream including 1 hour of video, 11 hours of audio, codebases with over 30,000 lines of code, or over 700,000 words.

It makes possible “provides “high-quality transcription and can be used to search audio and video content, such as using it to search, analyze, and answer questions across earnings calls or investor meetings.”

We’re seeing customers create entirely new use cases, including creating AI-powered customer service agents and academic tutors, analyzing large collections of complex financial documents, detecting gaps in documentation, and exploring entire codebases or data collections via natural language.

Similarly, Gemini Code Assist is using Gemini 1.5 Pro (in private preview).

There are a number of Gemini for Google Cloud announcements, including Gemini in BigQuery, Databases, Looker, Threat Intelligence, Security Operations, and Security Command Center. 

A new Premium tier of Chrome Enterprise provides advanced security capabilities, including more enterprise controls, security insights/reporting, context-aware access controls, and threat/data protection.

Google previewed Imagen 2.0’s new “text to live image” (or GIFs) capabilities. This will let marketing and creative teams create animations at 24 FPS, with a 360×640 pixel resolution, and 4-second duration. Google says it’s “adept at themes such as nature, food imagery, and animals,” with a “range of camera angles and motions.” All output will feature digital watermarks and safety filters.

More broadly, Imagen 2.0 is getting “advanced photo editing features” that allows for inpainting and outpainting.

Now generally available for Imagen 2 on Vertex AI, these features make it easy to remove unwanted elements in an image, add new elements, and expand the borders of the image to create a wider field of view.