Google just 'caught up big time' in the LLM race toward integrated system dominance

Since OpenAI took LLMs mainstream in 2022, the prevailing narrative suggested Google was playing catch-up in the generative AI race, despite originating much of the foundational technology powering the new wave of emerging tech. But according to Jonathan Hodges, VP of Data & AI for Google Cloud at global digital transformation giant Valtech, that perception is now outdated. Hodges argues that Google, through internal shifts and a massive suite of announcements made at I/O, has not only caught up but is now effectively leveraging its existing ecosystem to establish a dominant, integrated AI platform for companies of all sizes.

From 'catching up' to leading the charge: "Google is finally catching up," Hodges says. "Even though they originated much if the tech with transformers and early research, they didn't capitalize on it, so they got behind."

Hodges points to Google's I/O event as a clear turning point. "I think I/O is where everyone could finally see the full strategy," he says, pointing to strategic internal reorganizations, like bringing DeepMind more into the commercial side of the business.

Leveraging the ecosystem: A core part of Google's resurgent strength, Hodges believes, is its ability to weave AI through its vast existing infrastructure. "They've got the distribution via Android Mobile, web search, Workspace, all the Office tools where they can seamlessly 'bring this stuff to life.'"

The existing footprint offers a unique advantage, especially when it comes to data for personalization. "Think about what Google does with Google Maps, YouTube, and Android. Think about how much consumers want to personalize AI. Google is the company that has a lot more of that information than OpenAI or even Microsoft, and now that they're opening that up for their customers."

Enterprise search as a "massive" differentiator: One of the other significant advantages Hodges sees is Google's deep expertise in search being tailored for enterprise needs. "This enterprise search layer is so big. There's literally a verb for 'Google' that makes it synonymous with search. So bringing that capability and that expertise to enterprises is massive."

This isn't just about finding documents; it's about building a contextual understanding through knowledge graphs to empower AI models. Hodges explains that with platforms like Google's AgentSpace, clients get enterprise search for all their employees on "on day one", providing immediate value and a foundation for more advanced agent-based systems.

The power of a unified agent platform: Google's strategy extends to creating a cohesive environment for AI agents. "AgentSpace has built a way to bring that under that enterprise search umbrella. Google is the furthest ahead of being able to connect and bring in all the source permissions and security," Hodges notes. Crucially, this platform isn't a walled garden. "I can bring my Microsoft Agent Framework agents, or my agents I built in some open source frameworks like LangGraph. All these can now be deployed in the same environment, with the same agent engine, and the same permissions." This addresses a major enterprise pain point: "What organizations don't want is this proliferation of point solutions.

Driving real-world client value: For Valtech and its clients, these advancements are already translating into tangible benefits. Hodges mentions Valtech's own product, "Create" (CR8), which leverages Google's multimodal models (like Veo, Imagen, Lyria, Chirp, and Gemini). "We're working with folks like Bloomingdale's, L'Oréal, and other top brands. They want to be able to generate content and personalize to the point that even when you and I are looking at the same product, they may actually display it in a completely different way with completely different personalized description." This dynamic content generation for product display, marketing campaigns, and personalized messaging is becoming "part of the normal way of these retailers are going to do business."

Beyond retail, advancements in AI-assisted coding with tools like Jules (Google's always-on coding agent) and the AI Studio's "Build" tab are enabling Valtech to "bring costs down for our clients, shrink the time to delivery."

A widening competitive gap? Hodges believes Google's integrated approach and unique data assets create a formidable challenge for competitors. "I think that's going to be very hard for OpenAI and these others to close because they don't have the existing that literally billions of people literally use," he says. "They are going to tru to close that gap, but I'm struggling to figure out exactly how they are going to do that."