3 Takeaways from Google Cloud Next 2026

Estimated Reading Time: 1 minutes

Google Cloud Next 2026 wrapped up this week at Mandalay Bay, and I’ve got thoughts. Some of them are even about the announcements. But let’s start with the elephant, or rather, the 40,000 people crammed into a venue built for significantly fewer of them.

The Vibes: Herded Cattle in a Police State

Look, I’ve been to a lot of conferences. I understand the math of big-tech events. But whoever keeps signing off on Mandalay Bay as the venue for this conference owes every attendee an apology and a cocktail. The hallways were gridlocked; the session rooms filled up 20 minutes before start times; and I genuinely spent more time waiting in lines than in the sessions. The facilities were buckling under the load: bathrooms, food service, escalators, all of it.

And the security theater. My badge got scanned so many times I started to wonder if Google was training a model on my movement patterns. (They probably were. It probably launches next quarter.)

The Expo floor was its own tell. Walking the aisles, it was SIs and hardware vendors as far as the eye could see. The companies actually doing the work of turning enterprise data into something useful? Conspicuously absent. Read into that what you will, but I have a theory: Google wants to own that layer of the whole data and analytics stack now, and the ecosystem is reading the room.

Now, the Actual News

Despite the crowd-management disaster, there was real substance. Three themes stood out.

1. The Cross-Cloud Lakehouse

Google’s big data play this year is the Cross-Cloud Lakehouse, and it may be more than marketing, but time will tell. It’s powered by Google’s new Lightning Spark engine (Google claims it’s 2x faster than the fastest Spark) that sits on Apache Iceberg. This is Google positioning itself as the place where your data lives, regardless of where it was born.

The Iceberg bet is the tell. By leaning hard on the open-table format, Google is signaling it won’t play the proprietary lock-in game as it had in the past. The SaaS integrations through Iceberg are particularly interesting: your Salesforce, Workday, and ServiceNow data are becoming queryable as first-class tables in your lakehouse without the usual ETL gymnastics.

2. MCP, MCP, and More MCP

If you had “Google goes all-in on Model Context Protocol” on your 2026 bingo card, congratulations. Every Google service is now MCP-enabled by default. Every one. That’s not a feature, that’s a platform stance.

Even more, their agent tools have genuinely nice integration with external MCP servers. This is the right call.

3. Gemini Enterprise: Don’t Sleep on This

Gemini Enterprise was the most impressive thing at the show, full stop. Central to their announcement was their Gemini Enterprise Agent Platform. A few pieces worth calling out:

The agent marketplace. This is Google’s answer to the question every CIO is asking: “How do I actually deploy agents at scale without a boutique consulting engagement per use case?” They demonstrated several third-party agents that could be orchestrated via their agent framework.

The Agent Designer. This is where it got fun. Adding skills to an agent is genuinely streamlined, and the way it pulls in Google services like Maps through MCP is exactly the kind of first-party integration advantage Google should be pressing. This is their home-field moat.

The Agent Registry. A proper management console for your agent fleet. Boring, necessary, and the kind of thing enterprises will actually pay for.

The Planning Agent. The key piece, in my view. The orchestration layer that makes multi-step, multi-agent workflows actually reliable is the hard problem, and Google is taking it seriously.

A Few More Observations

The keynote was… scripted. Heavily. Painfully. Every customer testimonial was vague, fulsome, and interchangeable: the kind of quotes that make you wonder if anyone at those companies has actually put the product into production or if PR wrote them over coffee. “Google Cloud has transformed how we think about data.” Cool, thanks for sharing.

One thing I did appreciate: Wiz, the $32B acquisition that closed in March, showed up in the keynote demos as Wiz. Not rebranded. Not “Google Cloud Security Suite.” Wiz. That’s a smart brand decision; Wiz has earned its name in the market, and burying it under Google branding would destroy value.

Bottom Line

Strip away the crowd-management disaster and the corporate-theater keynote, and Google Cloud Next 2026 was a serious signal. Google is making hard, specific bets: open table formats, MCP everywhere, agents as the primary enterprise interface, and a claim to own the enterprise catalog and knowledge graph layer. Some of these bets will land. Some won’t. But this isn’t a company coasting; it’s a company playing offense.

Now, if they’d just rent a bigger venue next year. Or, I don’t know, thin the herd by making the keynote testimonials pass a Turing test.

SHARE
Guide: How to Choose a Semantic Layer
The Ultimate Guide to Choosing a Semantic Layer

See AtScale in Action

Schedule a Live Demo Today