Pushing What’s Possible: Reliving Google Cloud Next 2025

Google Cloud Next 2025 just hit different. There was a vibrant, fun, and exploratory energy that came across in every aspect of the event. From the colorful branding to the relaxed delivery and unconventional venues, it felt more like a festival of ideas than your traditional tech conference.

But don’t let the chill vibes fool you. Beneath the relaxed surface was an uncompromising push for progress underpinned by a commitment to technical excellence. Gemini 2.5 is now ranked #1 in a LLM leaderboard voted for by over a million users. More than 200 new product announcements across the three-day summit only highlighted the pace of innovation.

Despite their playful approach, it’s clear that Google Cloud is not here to play. Here are my stand-out moments from the event.

Generative AI goes to Hollywood

I have to start by addressing the elephant in the room… or should I say the cowardly lion?

The Google Cloud partner community had the privilege of being the very first audience to witness a demonstration of their multimedia generative AI capabilities. They played a scene from the classic film ‘The Wizard of Oz’ expanded to fit the highest resolution canvas in the world — the 16K, panoramic, LED screen at the Las Vegas Sphere.

Reimagining the film for this format involved months of collaboration between thousands of developers and creatives from Google Cloud, DeepMind, Magnopus, Warner Bros., and Sphere Entertainment Co. As well as upscaling the low resolution 1939 original in astounding detail, the team also carried out a process called ‘outpaining,’ extending frames to insert characters like the Tin Man into shots they had not appeared in previously.

This reinvention of an ageless classic demonstrated how AI can redefine what’s possible in visual storytelling while also preserving period charm, revitalizing an old film for a new generation.

Google is now the only company that offers generative AI models for video, image, speech, and music. This is a key differentiator for any business producing content, running marketing campaigns, or creating digital experiences. I expect its impact to be felt far beyond the media and entertainment industry.

Google Cloud ushers in the age of inference

Over the past year, we’ve seen Google’s AI development pull away from the pack. Google Cloud Next 2025 only extended that lead.

As well as raising the bar for multimedia AI, they’re also setting the standard for technical performance. They drove that point home with the announcement of Ironwood, Google Cloud’s seventh-generation Tensor Processing Unit (TPU), a custom silicon chip boasting five times more computing power and six times more memory capacity than the previous version.

Ironwood is designed to power thinking, inferring AI models at scale. Its launch reflects a move from responsive AI models that compile and deliver information, towards proactive generation, where AI will return insights — not just outputs.

AI models like Gemini 2.5 and AlphaFold run on TPUs, so this leap forward will enable developers to meet the rise in computing demand with confidence, and power the next generation of AI breakthroughs. And judging by the reaction in the room, the market is ready.

Bringing AI to every employee with Agentspace

Also high on the agenda were updates to Google Agentspace, a marketplace for powerful agents built on Google’s latest foundation models that put enterprise knowledge in the hands of employees. With new features including Chrome integration and a no-code Agent Assembler, it’s now easier than ever for customers to create, find, and deploy agents from their chosen vendors, which can then connect instantly to their own AI models.

Imagine downloading a Salesforce agent that could talk fluently to your AI model from day one, allowing it to query and analyze your CRM data to return higher-value insights and advanced reporting. That’s the promise of Agentspace, and I think it will be game-changing for multi-source AI workflows.

A shared vision for a multi-cloud future

Nasuni has been a Google Cloud Premier Partner since 2021, and I was thrilled to be invited to contribute to the Google Cloud Transform Studio — a showcase of inspiring cloud and AI use cases, driven by customer success stories.

But our alignment with Google goes beyond credentials and installations. One of the clearest messages from the event was Google’s commitment to multi-cloud AI. Rather than locking customers into Google Cloud, they are focused on making AI tools that can travel to where your data is — whichever cloud they live on.

This multi-cloud AI philosophy is well matched with Nasuni’s ability to support and manage multi-cloud architectures through a single global file system. Out of the hyperscalers, only Google champions this approach and it represents a powerful shared vision between our two companies. From our own conversations with customers at the conference, it’s clear that multi-cloud flexibility is a top priority for them too.

The yellow brick road of cloud and AI possibility

I’ve seen technology solve incredibly difficult and complex problems. But what we witnessed at Google Cloud Next 2025 felt different. This was technology breaking down the door to what was previously thought to be impossible.

And on the other side of that door?

A yellow brick road of possibility that we’re excited to travel with our partners, customers, and peers. It’s clear to me that the things that we will accomplish with AI tomorrow are beyond what we can imagine today. I’m looking forward to the exciting heights the Nasuni and Google Cloud partnership will reach this year.