Google Cloud AI Announcements

By Patrick Moorhead - December 20, 2023

The Six Five team discusses the Google Cloud AI Announcements.

If you are interested in watching the full episode you can check it out here.

Disclaimer: The Six Five Webcast is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors and we ask that you do not treat us as such.

Transcript:

Patrick Moorhead: Google Cloud’s AI announcement, and note, it’s Google Cloud, not Google, because we did see some announcements on the prior week about Gemini and stuff like that, but what did Thomas Kurian have to say that was incremental here?

Daniel Newman: Well, it’s really about the integrations. What we heard about a week ago was Gemini, what we heard about this week is how it’s going to be used in Vertex Pat. How it’s going to be part of Google AI Studio, which is basically the enabler for people to be building on Google Cloud using Gemini. We’re hearing about this. You kind of just go through the last year, you’ve heard about Azure Cloud services with OpenAI. You’ve heard about everything that AWS is doing with their offerings, and very much an open approach. Well, Google is very centric to its models. Google’s approach was always about building a full stack capability, and this week he was basically focused on the integrations and being seen as clearly caught up in the market to OpenAI. By the way, are you flashing those up there?

Patrick Moorhead: Yeah, yeah. We’ve got some people coming in from Twitter. We’ve got some people coming in from LinkedIn just to say hi. That’s all. Hello, thanks for joining.

Daniel Newman: You hadn’t done that in the past. Our producers are getting cooler and cooler as it goes on.

Patrick Moorhead: Dan, we keep adding capabilities to the Six Five. And who knows, we might go live at CES 2024. Oh, we are. Okay.

Daniel Newman: So let’s kind of run this back. You’ve got AI HyperComputer, which is going to be a train and serve generative AI models. You’ve got generative AI support and Vertex with Gemini now. You do have, what I did like by the way, is I did talk about how they’re building this conducive all Google, but you did hear them talk about also being open. And this is something I think is very important. They mentioned anthropic, they mentioned Cohere. The company did do a good job. I thought yesterday, was it yesterday? It was two days ago. Thomas did mention customers, which to me is always very important that are going to be building and engaging and utilizing the overall Google Cloud stack. Pat, I did say something earlier, it was interesting. I heard the same stat from two company about unicorns. Google, 80% of unicorns are building in Google Cloud.

Ironically, 80% of unicorns are also building in the US. By the way, if there’s ever been a proof point for multi-cloud, do you not feel like that was the ultimate proof point is the answer is probably yes. You have 80% of unicorns are using multi-cloud. So what else? You have Duet Pat. So Duet, the assistant AI agents were announced. So you go up and down the stack, it really felt to me like an extension, not anything so much new, and I’d love to get your take on this Pat, but everything we heard from Thomas was my expectation, meaning the first week was about Google holistically coming to market and saying, “Here’s what we’re building as an LLM, here’s our strategy, here’s our silicon.” You and I have put some thoughts on, you wrote a great tweet on the TPU 5, doing comparatives. But overall what I expect now, it has to make its way over into the sandbox, and Google Cloud is the sandbox.

If you’re a company and you want to build applications and develop capabilities using generative AI, they have to make these models available and accessible and easy to use. So that’s really what Thomas was focused on, in my opinion, Pat. So it was more of an extension to me of the news than it was necessarily anything new. I don’t know. What did you think? Did you think it was, I guess let me ask you that. Was there anything surprising to you? To me, this all felt very expected, but I was happy to see how quickly they moved to make these things available in Google Cloud.

Patrick Moorhead: So most of this was to be expected. I did appreciate Google Cloud CEO, Thomas Kurian doing a live Q and A with the analysts. I showed up late, but I was appreciative of the time he spent. I thought it was time well spent because quite frankly, I’ll call it the consumer version of Gemini that came out didn’t land very well. There were some really good things about it. I mean, quite frankly, they’re talking about a future service that’s 25% better than what’s already in market with GPT-4. And there were some questions about the demos and the duck, the Blue Duck. I mean I think that’s what it was. And so it didn’t land very well and it also didn’t seem like there was a lot of prep time for press and analysts running up.

Daniel Newman: We talked about this last week. I think it’s actually important. What is the openness in the world to companies showing capabilities versus what’s expected? Is there any longer an acceptance Pat for companies to sort of do those, we’re in this world now where everybody’s pulling apart the threads on everything. And companies used to, I remember 20 years ago in the tech industry, there was someone behind the curtain showing the devil of what was coming. And now it’s like, is there latitude for that?

Patrick Moorhead: So Dan, I think it’s context in history. If you remember what happened in the first Google Bard introduction, it was so bad that they pulled the video down on YouTube. I was trying to look for it, because was done out of France and it was on a different time zone and I couldn’t find it. They pulled it down it was so bad. So then I think people were on high alert, I mean Apple for example, has a decent track record showing demos of what’s going to be, and then the experience ends up being that. A little bit of scrutiny but not a lot.

So context in history came in. So I think everybody was on alert for this. And I think it’s super important always to show this. These big huge companies are getting away from disclaimers at the bottom of let’s say a video, and even a claim. And even Apple is very poor at this when it makes performance claims, it doesn’t actually disclaim in the video. It disclaims maybe on their website afterwards at the very bottom. So I do think that we haven’t shifted as an industry yet to some we don’t believe anything and we can’t show vision. I really think it’s context in history, but getting back-

Daniel Newman: Thanks for that. I was just really thinking, I didn’t mean to derail you, I just really wanted to, are we entering the era where we’re going to scrutinize every demo we see now? Or do you think to your point about context, is it because they botched that first one that now every demo they do is going to get a level of scrutiny? Because Apple, like you said, they get away with everything. Nobody asks anything. You and I are the only two good people on the planet that ask questions of what Apple shows.

Patrick Moorhead: Yeah, and I’m pretty sure that that was done by Google DeepMind, which is research which you would expect this long pull. So by the way, Dan, I do like going down these ancillary topics, related topics. I don’t think we should be completely programmed to where we can’t talk about something that we think our viewers are interested in. So this was the, oh by the way, all that stuff that you saw with Gemini and TPU 5, it is going to be available to enterprises too. Like you said with Vertex, with Duet. I think the important point that was that Google Cloud CEO Thomas Kurian came out and made that message because think about if you’re an enterprise user and you’re seeing the Blue Duck, you’re seeing all this stuff. I mean I think the first article showed up on Engadget, right? Wait a second, is this enterprise-grade technology or not, and how should I relate to it?

There was something very provocative that I just want to let you know I need to do more research on. But it was running data and AI anywhere, with essentially Google’s distributed cloud. Google’s distributed cloud is essentially a hybrid offering that runs on-prem at a data center, it runs at a colo. Again, I’m looking for the company that can give me a reasonable layer cake diagram of exactly how the pieces fit together, how on-prem, how Edge and how public cloud all fit together. So I need to do the double click, but essentially what I’m reading is is that they’re putting Xon, and by the way, Google Cloud was the original hybrid offering that the name escapes me. They got away-

Daniel Newman: Anthos.

Patrick Moorhead: Sorry?

Daniel Newman: Anthos.

Patrick Moorhead: Exactly. Thank you very much. But they’re putting Xons out there, I don’t think they’re putting TPU out there, but I need to do the double click on there. This reminds me a little bit of Oracle Cloud@Customer. You have one, two, three, four, five, six seven, eight different types of implementations. And by the way, six of them are new. Modular data center, scalable multi-rack, tactical appliance, HA multi-server, high available. So anyways, I needed to double click on that. I didn’t feel like there was a lot of time spent on that and I didn’t get pre-briefed. So I’m going to ask Matt and the team to dive in and see what’s going on there.

Patrick Moorhead

Patrick founded the firm based on his real-world world technology experiences with the understanding of what he wasn’t getting from analysts and consultants. Ten years later, Patrick is ranked #1 among technology industry analysts in terms of “power” (ARInsights)  in “press citations” (Apollo Research). Moorhead is a contributor at Forbes and frequently appears on CNBC. He is a broad-based analyst covering a wide variety of topics including the cloud, enterprise SaaS, collaboration, client computing, and semiconductors. He has 30 years of experience including 15 years of executive experience at high tech companies (NCR, AT&T, Compaq, now HP, and AMD) leading strategy, product management, product marketing, and corporate marketing, including three industry board appointments.