At Google Next 2025, RedMonk’s James Governor highlights a major shift: Google’s swagger is back. Amid fierce AI competition, Google stands out as the only hyperscaler leading with its own model—Gemini 2.5. While rivals rely on third-party models like ChatGPT and Claude, Google is doubling down on native innovation. The result? A more precise, developer-friendly experience that reflects growing confidence in its enterprise AI strategy.
Transcript
Hi, I’m James Governor, co founder of RedMonk, the industry analyst company that focuses on developer-led tech adoption. So I’m here at Google Next 2025, and I’ve got some quick takes, some thoughts about the event. And what have I come away with? Well, here’s the thing. Now, obviously, with the incredible rise, the explosion of innovation that we’ve seen around generative AI, we saw the incumbents, the Hyperscalers, perhaps on the back foot just a little bit. I would say that what I’ve seen in 2025 from Google is the swagger is back. This started from a demo that they’re doing some work around the Wizard of Oz in the Sphere here in Las Vegas. Just seeing when you’ve got Sundar there, and he is very much happy about the situation. You could see that Thomas Kurian in the same. One of the questions one of the analysts asked Thomas at the event was, what was a little smile on his face and what was that about? I think the truth is that there is a confidence, certainly in this enterprise business and in this developer-led business around generative AI. If we get specific about that, the models that we’ve seen, pretty clearly, 2024 was, well, 2023 was all about ChatGPT.
It was all about Open AI. 2024, Claude was smashing it. So Anthropic did a great job. Developers were loving Claude 3. 5. But what’s pretty clear right now is there is only one hyperscaler that is leading with its own model, and Gemini 2. 5 from Google is doing some really good work. Claude can be a little bit almost verbose. You have to keep on querying it to narrow down in order to get the prompting right so that the prompts behave in the way that you want. It’s still very effective, but sometimes it almost wants to do too much. If there’s one of the things that we’ve done in software engineering over the past, well, two decades, it’s like make smaller changes. Don’t change everything all at once. Otherwise, you might mess things up. Small changes can be good, and you want to be very clear about specification. One of the things that we are seeing with Gemini 2. 5 is actually it just does what you tell it. I think that’s a huge opportunity for Google to move forward. So they’re very much what we’ve seen in the last couple of days.
It’s Gemini first, 2. 5 first. Everyone else, they keep wanting to have to talk about Claude is going to be in there because that’s where the work is being done. So we’re in a world where we’ve got a lot of interesting innovation around code assistance. I think that for me is probably the one takeaway, which is that Google, in a time when we have to move quickly because the competition is so fierce, while you’ve got Microsoft, they’re reliant on ChatGPT, you’ve got Amazon, they’re very much pushing the multi-model thing. Google is happy to do multi-model, but they’re the ones that are pushing their own model first. And that’s what we’ve seen. I think that’s where the swagger comes from.
