Google has just announced that their AI overviews, right?
So when you do a Google search, and there's a little AI summary at the top, those have reached 2 billion monthly users.
Now, this isn't really a shocker, because basically, people aren't opting in or opting out.
Google is just putting those at the top of their Google search results.
They have 2 billion users on Google.
They're giving those to everyone.
So now there's 2 billion monthly users.
I think what was more shocking to me was when they announced how many people were actually going and using Gemini, the standalone app or website, so people have to go out of their way to use it.
And with all of this seemingly, you know, like, oh, positive news for AI and Google, they just signed a deal with OpenAI that some people are saying is pretty, is called, feels just like the end of Google, the same way Google signed a similar deal with Yahoo back in the day.
So I'm going to be breaking down what's going on, the new deal that OpenAI and Google have basically partnered and signed aligned together what's happening with that, and if it really is the end of Google.
Before we get into all of that, if you want to try any of the models that I talk about on this show, including everything from OpenAI and everything from Google, Gemini, all of that, I would love for you to try out AIbox .ai.
This is my own software startup.
It's currently in beta, and we have the top 40 AI models all on one platform that you pay one subscription and you can try out and use everything.
So you don't have to have subscriptions to every single platform.
We have a whole bunch of cool cool features in there that you won't get anywhere else one of those being a media storage folder so every piece of media that you generate you're able to go in and click on i had this problem with chat gpt where i would like generate a picture and then like a month later i'm like dang i really need to go like look and i want to like use basically like a thumbnail for a video or something and replicate it or make a spin on it or something like that and i couldn't find what thread it was in it was so far back and this isn't just for thumbnails but all sorts of files
that that you would generate.
We've created something called a media storage where basically you can go, you can see the prompt that you use to generate an image.
You can also then go and view directly the chat that you were in when you generated that image.
So if you had previous contacts or messages, you go see all of that.
You can switch between different AI models to generate.
So you could give it one prompt, switch between different AI models, and you compare all of your prompts side by side from the different companies.
So a whole bunch of cool features.
You can go check check it out, AIbox .ai.
But let's get into what Google has announced and the partnership that they have going on right now with OpenAI.
So the first thing I wanted to share is that Sundar Pichai has said that he is quote unquote, very excited about a Google Cloud OpenAI partnership.
Now, basically what's happening here is that OpenAI has made a very deep partnership with Microsoft. They've spent a ton of money on Microsoft Azure for all of their cloud computing, right?
That's kind of the premise of their partnership.
Microsoft gave them billions of dollars and they're using Microsoft Azure.
There came a problem, basically.
There was a lot of tension with OpenAI and Microsoft where they didn't have enough compute.
Microsoft Azure couldn't give them more, basically, because they're pretty strapped with all of their customers.
They're building it out, but everybody's building out.
We have a moment right now where OpenAI is building out a $500 billion data center project with Oracle and with SoftBank.
And I think they literally just don't have enough.
So OpenAI has got in a position where basically they have to go and shop around and find other people to use.
Now, the biggest companies are Microsoft is your Google Cloud and Amazon web servers.
Their arch nemesis and rival Anthropic is really heavily embedded with AWS, right?
AWS has given them more than $4 billion.
So very tight partnership there.
there. And also, to be fair, Google Cloud gave Anthropic $300 million, but that's smaller.
And that was kind of, I think, like over a year ago.
So basically, they've come to Google Cloud to help run and get more compute, more access to more compute.
So this is, you know, theoretically, this is great for Google, because it's more money to their cloud platform, which, by the way, is growing very fast. But some people are saying, you know, they're spending, Google is spending $10 billion to, basically on CapEx, to try to catch up, basically exclusively on AI, to try to catch up with OpenAI.
People are saying, you know, they view this as such an existential threat to the company that they need to spend $10 billion to catch up on this quote -unquote AI race.
So that's roughly two and a half, it's been roughly two and a half years since ChatGPT came out.
And Google is basically focusing all its attention there.
I think we're seeing that ChatGPT is showing no signs of stopping.
People have completely replaced, myself included, doing Google searches generally with just asking AI.
And so Google knows that if they don't create a very strong competitor to chat GPT, no one's going to use Google, many people will not use Google, they'll just go to chat GPT, or Gemini, or grok, or anthropic, or any other like AI model, basically, it's just a better response.
And so that's why they put so much time and energy into Gemini, I think.
But it's kind of an interesting point.
Because right now, they're giving all of this compute to OpenAI to train better models, because OpenAI has basically given a case study and proven that if they have more compute, they can make their models cheaper and better.
Right now, because the compute is so expensive, OpenAI is like, yeah, if we spent $1 ,000 for every single question that you ask ChatGPT, our results would be like 50 % better, more accurate, PhD level, math, reasoning, science, like the results are just inarguably better on all the benchmarks, marks, but it's so expensive.
And so how they can bring that cost down is by getting access to more compute and training the models more.
But in order to get the price of compute down, they have to do, you know, economies of scale.
So that's why they're doing these, you know, $500 billion deal with Oracle and SoftBank to build out more data centers.
But in the meantime, they need to use other people.
So this is kind of where this deal came with Google.
And it kind of like in a weird way sort of leaked, CNBC wrote an article on it.
But basically OpenAI kind of added, people are they're saying they quietly added Google Cloud to their public list of suppliers.
And people were like, wait a second, Google Cloud's one of your suppliers?
You're a very deep competitor with Google Gemini.
And so this is basically who they're using alongside Microsoft and Oracle.
So they have Microsoft, Oracle, now Google Cloud.
Google Cloud revenue, by the way, in case you're interested on where that's at these days, in 2025, they were at, so this year they're at $13 .6 billion this is the second quarter of 2025 and that's up from 10 .3 billion in the same quarter last year right so they're up more than three billion dollars um and largely that's all due to ai companies and the products that they're offering ai companies to train with google cloud and so obviously this deal that micro that opening is doing with them is going to help increase that but also it increases the competition because opening is going to get that much
better it's going to get that much cheaper and gemini is going to have to spend more so it's kind of of interesting where Google's at a point where they're spending $10 billion on CapEx to keep up with OpenAI.
Meanwhile, OpenAI is also doing deals with them that like it's giving them money, but like, you almost wonder if the money that they're giving Google is less like it's worth less than the money they're gonna have to spend to keep up with OpenAI.
But in any case, OpenAI is gonna find a way to get that compute.
So there's, it's not really like they want to, and they also don't want to be like anti -competitive, I'm assuming and get into bad graces of any sort of regulators if they were going to like stifle OpenAI for anyway.
So OpenAI is not the only big lab using Google Cloud.
We have Ilias Susquehanna's safe super intelligence, Fifi Lee's world labs, Anthropic is using them.
And so a lot of people are using this.
And this is basically because they have a huge supply of NVIDIA GPU chips, and they have a bunch of in -house TPU chips.
Both of these are very good for training AI models.
It's a great partner.
And so, you know, I think NVIDIA has been doing a lot with them.
And this is obviously just kind of a point where open AI has these tensions with Microsoft and they got to make this play.
So this is an interesting point.
Open AI and, or I mean, Google is obviously growing very fast with everything they have with AI.
Basically, like I think the headline everyone has heard is like, oh my gosh, 2 billion monthly users for their AI summaries.
I was more impressed by the fact that Google Gemini right now has reached 450 million monthly active users.
You actually have to go to the Gemini website or the Gemini app, I think for that to count when you're talking about that.
So that's a pretty impressive number.
It's half a billion.
So a quarter of the Google users getting the AI results are also going and using Gemini one way or another.
And perhaps maybe they're sneaking in some other things like use cases in Gmail and stuff.
stuff. But in any case, it's a ton of users.
I think it's very interesting.
I'll be curious to see how long this relationship with OpenAI and Google lasts as they're deathly competitors right now.
And both of the companies are basically battling for the same people looking for the same questions.
So it's gonna be very interesting.
Hey, if you enjoyed the podcast today, the number one thing that I would appreciate is if you could leave this video a like and subscribe over on on YouTube or a rating and review over on Spotify or Apple helps the podcast a ton to get found by more people and to get into the algorithm.
Thanks so much for tuning in and make sure to go check out AI box .ai.
If you want to try all the top 40 AI models for one price and get a whole bunch of cool features like comparing one prompt side by side with a whole bunch of AI models, you can ask, you know, one AI model to generate you a picture of a red panda, you can ask ideogram, open AI, AI, Flux, and you can compare all of the images side by side or all of the text outputs side by side.
You get it to rewrite you an email or write you a blog.
You can look at all the blogs side by side.
See what AI model is the best for your task.
Thanks so much for tuning in to the podcast. I'll catch you next time.
