Google Cloud Next 2025: Targeting easy AI
Las Vegas has been taken over by Google Cloud this week, with Google Cloud Next 2025 in full flow.
There have already been a number of headline grabbing announcements: we’ve seen the launch of a powerful new AI accelerator chip by the name of Ironwood, huge updates to Google Cloud’s security capabilities, and new agentic AI development kits for enterprises. But there’s been much, much more on top.
In this episode, Ross Kelly, news and analysis editor at ITPro, speaks to Rory about Google Cloud’s messaging at its annual conference and what the mood on the ground is at Google Cloud Next 2025.
Highlights
“The idea here is that if you are deploying an AI agent in your environment, it should be able to communicate with other AI agents, regardless of which vendor built them or which framework they’re built on.”
“So if you’re a really large enterprise and you’ve heard those figures I’ve already given and you’re thinking, ‘well, we’re going to need more than that for presumably AI training, but also it could be for inference on a really, really large scale. You can connect up to hundreds of thousands of Ironwood chips together using the pathways runtime.”
“That interconnectivity is quite quickly building up into this new ecosystem for the cloud. It ties in interestingly, that theme, to one more Gemini announcement actually, that Google Cloud made at the conference, which is that Gemini will now be available on Google Distributed Cloud, which means that if you are an organization that is looking to you can run Gemini on prem so in in your own data center, including in an air gapped environment.”
Subscribe
TOPICS
Source link
