It feels almost like yesterday that Google was falling behind everyone on AI. While ChatGPT brought in hundreds of millions of users, its chatbot, Bard, couldn’t stop hallucinating. Google search was telling people to eat rocks, and they even had to roll back their image model after it generated some controversial stuff. To put it mildly, things were not looking good for Google.
But fast forward to today, and it’s a very different story. Open AI declaring code red as the AI race intensifies. Shares of Alphabet are hitting all-time highs today, fueled by its launch of the new AI model, Gemini 3, and an agreement to help establish Google’s chips as a rival to the industry’s bestselling AI accelerator made by Nvidia.
In November 2025, Google released Gemini 3, its most powerful language model yet, and it was so much better than GPT5 that in response, OpenAI declared a code red. That’s what’s been making most of the headlines lately.
From AI Embarrassment to Market Panic:
But here’s the thing. Gemini 3 isn’t just a threat to open AI. Google is now competing with Nvidia, Oracle, Microsoft, Meta, AMD, basically every AI company you could name. And they’re doing it in a way that no other company possibly could. The more I research this, the more it starts to look like no matter what happens in AI going forward, Google is going to win.
Rewinding to Google’s Messy AI Strategy:
But to understand why, you have to go back to when Google’s AI strategy was still a complete mess. Funny thing is, it wasn’t even that long ago.
The Antitrust Trial That Changed Everything:
Let’s go back to 2020, before all of this AI stuff happened. At the time, Google was in the middle of one of the most important trials in its history. The US government had accused them of illegally monopolizing the search advertising market, and the jury actually found them guilty. It even looked like Google would be forced to spin off Chrome and Android into their own separate companies. But that never happened. You see, when this case began, there was really no credible threat to Google’s search monopoly.
But then, in November 2022, OpenAI released ChatGPT as a small experiment. We all know what happens next. Two months later, ChatGPT had 100 million users, and the entire tech landscape was pivoting to focus on AI. Now, at the time, everyone was saying this was a massive threat to Google. The company’s main profit driver was and still is search ads. They make more than a hundred billion dollars from it every year.
But when compared to a Google search, most of the time, ChatGPT was just better. Its responses were more nuanced and personalized, and since you could just talk to it instead of trying to use the right keywords, it was much more intuitive all around. As you can imagine, this put Google in a kind of an awkward spot. On the one hand, if they did build an AI product, it would compete directly with search and threaten their biggest source of revenue. But on the other hand, they couldn’t afford to build one.
Why Google Could Not Ignore AI Anymore:
AI was the future, and they had to be a part of it. So, by 2024, when the judge overseeing Google’s case had to decide what to do, AI had completely changed the nature of search. Ultimately, the judge decided against breaking them up because they thought about it. Yeah, Google was guilty of monopolizing search, but now that the entire business model had competition, Google had to do something about it.
Now, the truth is, they had been trying. It just didn’t go well. A few months after ChatGPT came out, Google released Bard, their own chatbot. But in its demo, it hallucinated a basic fact in front of the entire world, which made Google lose over a hundred billion dollars in value just that day. A year later, they rebranded Bard to Gemini, which now came with an image generation model. For some reason, though, it insisted on making everyone, even these guys, racially diverse.
Then, when Google introduced AI overviews into search, they were telling people to eat rocks and add glue to their pizza, even that smoking while pregnant was perfectly healthy. I mean, this wasn’t good. For a company as renowned as Google, it was frankly embarrassing. People were arguing that they were simply too big and too bureaucratic to move fast on AI.
The Sudden Shift After the Case Ended:
But looking back, it’s almost as if they were waiting. Ever since Google’s case ended, they’ve been releasing AI products left and right. To VO3 and Nano Banana, their video and image generation tools are now, without a doubt, the best in the market. And Gemini 3 was literally so much better than GPT5 that OpenAI declared a code red.
But this isn’t even just about chatbots anymore. I mean, sure, OpenAI is in crisis mode. But you know who’s also under threat? NVIDIA, Oracle, Meta, Microsoft, AMD, Anthropic, Apple. Google is now competing with all of them. And they’re in a position to win. Not just the individual battles, but the entire war for the AI industry itself. You can see what I mean when you zoom out to see the entire AI stack.
Basically, all of the infrastructure behind a product like Gemini or ChatGPT. Fundamentally, there are five main layers to it. First is hardware, the physical device you use to access AI. Then there’s the product. This is like your ChatGPT app or Google’s AI overviews. Then, behind those interfaces, there’s the foundational model. These are your GPT5s or your Gemini 3s, which is the actual technology.
Next, those models were running inside data centers, big warehouses filled with advanced chips that do all the computation. And finally, there’s the design of those chips, which is the last layer of the stack. When you look at the tech landscape right now, you can clearly see where those companies I mentioned fit in. Meta and Apple make hardware. OpenAI and Anthropic make products and foundational models. Oracle and Microsoft run data centers. And Nvidia and AMD design the chips.
The difference with Google is that they don’t specialize in any one layer of the stack. They cover everything. And this vertically integrated approach is what makes them so dangerous. Right now, we are just barely starting to see Google’s absolute dominance in AI. But to truly grasp just how superior their position is, we have to take their strategy apart layer by layer and see how everything fits together. So, let’s go through it from the bottom to the top.
How Google Is Challenging Nvidia, Microsoft, and OpenAI:
Starting with chip design, the company you know for this is Nvidia. They’re now the world’s most valuable company because pretty much every single one of the world’s most advanced large language models is trained and run on chips that they designed. Or at least that was until Gemini 3. See, Google also designs chips. They just do it a little differently than Nvidia. For starters, Nvidia’s main chips are GPUs, graphics processing units. But Google designs TPUs where the T stands for tensor.
Without getting too much into the science here, GPUs and TPUs are both just computer chips. Fundamentally, what they do is math. The difference is that GPUs are great at a bunch of different kinds of math, whereas TPUs are specifically built to do tensor math, the kind that powers AI. What was groundbreaking about Gemini 3 was that not only was it better than every other model at the time, but it was also trained exclusively with Google’s TPUs. Gemini 3 is proof that Nvidia is no longer the only company making AI chips at the cutting edge, and this is a massive threat to them. Google’s TPUs are at least on par with theirs. And they literally just announced that they plan to double production by 2028.
The Unfortunate Scenario For Nvidia:
But unfortunately for Nvidia, it gets even worse. Up until very recently, Google was only building TPUs so that they could rent them out through Google Cloud. But now they’re competing even more directly with Nvidia by selling them outright. They just cut a deal to sell 1 million TPUs to Anthropic. There are also rumors that they may be selling them to Meta and OpenAI, too. Those are some of Nvidia’s biggest customers.
Google is literally breaking its monopoly, which is why, naturally, its stocks are going in opposite directions. Now, this is already a fantastic position for Google to be in. They’re joining the market that made Nvidia the world’s most valuable company. But remember, this is only the first layer of the AI stack.
So, let’s see how it all connects. The second layer is data centers. In case you didn’t know, this is the main way companies like Oracle and Microsoft are playing the AI game. LLMs need a lot of computation. So, what they do is build entire warehouses usually filled with Nvidia chips, and then they rent them out to companies like OpenAI to train and run their models.
Again, Google does this too. It’s their cloud service. But the difference is they make their own chips, and that makes it a fundamentally different business. Think about it. Microsoft and Oracle are in a pretty tough spot. Not only do they have to pay the markup for Nvidia’s chips, but between construction, equipment, and energy costs, their margins become razor-thin.
Google’s Own Designs:
On the other hand, Google’s situation is completely different because they design their own chips. They don’t have to pay Nvidia. They just eat the cost. But more importantly, they can custom-design their chips, their hardware, and their infrastructure to work together perfectly. And just think about what happens when you do that. Everything becomes more efficient, especially when it comes to energy. The result is that Google can run its cloud service at a fraction of the cost of what Microsoft or Oracle can. And it works out beautifully. You know, back when ChatGPT came out, everyone was talking about how much of a threat AI was to search.
But what I think we all underestimated was how beneficial it would be to Google Cloud. Right now, their sales are growing at an annual rate of 30%, and it’s now nearly 20% of Google’s revenue. And when it actually comes to AI, well, they figure that one out as well. The next layer of the AI stack is foundational models like Gemini 3, GPD5, also VO3, Sora, that kind of stuff. Now, at the start, OpenAI had a clear lead here. GPT 3.5, the first model ChatGPT shipped with, was revolutionary. And by the time they released GPT4, everyone else was still catching up to them. But GPT5 was different.
The Breakthrough Moment:
To a lot of people, it felt less like a breakthrough and more like a minor upgrade. Plus, by this point, OpenAI’s lead was gone. Google’s video and image generation models were already much better than OpenAI’s. And in November 2025, they released Gemini 3, which outperformed GT5 on basically every benchmark you could imagine. Now, OpenAI has just released GPT 5.2 this week, as I’m recording, and from what I’m reading, it is competitive with Gemini. But ultimately, if you zoom out, both OpenAI and Google are going to continue improving their models long into the future.
The way I see it, whatever benchmarks they break on any given Tuesday isn’t really what’s important. What matters is this. This is a graph of the Gemini app’s monthly active users. And now you see why OpenAI declared a code red. The real leverage in AI is in owning the product. And this brings us to the next layer for this new product category, which I guess is called a chatbot. Open AAI and Google are really the two most viable competitors.
Only nerds use Claude or Deepseek. Right now, ChatGPT is getting close to 900 million weekly active users, which is freaking insane. Gemini, by comparison, has 650 monthly active users, but it’s growing very fast. See, these apps are still in the process of going completely mainstream, and both Google and OpenAI stand to gain a lot from ads whenever they decide to turn them on. The only reason they haven’t done this yet is that the way these things work is they become better over time as they get to know you. And once they do, it’s very hard to switch. That’s why Google and OpenAI are doing everything they can to steal each other’s users. If they don’t do it now, it’s going to be too late.
The Ecosystem Advantage No AI Company Can Match:
Now, you might very well argue that dethroning ChatGPT will be immensely difficult, even for Google. And you might be right. But again, when you zoom out to see the bigger picture, Google still crushes OpenAI on the product layer. Their weakness and Google’s biggest strength is that chatbots aren’t the only product that matters. OpenAI has ChatGPT, GPT, and Sora. But apart from Gemini, Google has Chrome, Gmail, Drive, Docs, Sheets, Slides, Search, Maps, Calendar, Meet, Photos, YouTube, Android, and the list goes on.
This ecosystem is something no other company on Earth can compete with. I mean, maybe Apple if they got their act together. And Google, in some form or another, is making AI a part of it. Once you’re using it, which I mean, if you use Gmail or Google Docs, you will, switching to anything else is going to become incredibly difficult. And so, just to put the final nail in the coffin, let’s look at the last layer, hardware.
The main way most people access AI right now is through our phones. And you guessed it, Google does this too. They have the Pixel line, which is getting more and more popular every year. But apart from this, they’re also going to be releasing their own AI-powered glasses soon. And look, I have my doubts, but if the tech bros are to be believed, this is going to be the next big computing platform. If it is, you better believe Google is going to be leading the charge.
Conclusion
So that’s the whole AI stack, and you can see now how Google has positioned itself at every level. Right now, we’re just getting our first glimpse into Google’s absolute dominance in AI. And it all stems from the fact that they’re completely vertically integrated. It’s a structural advantage that I just don’t see any other company replicating. Plus, let’s not forget that Google is Google. They’ve been a machine learning company since the 1990s. They have one of the best AI research teams on the planet.
And of course, they have that Google money. Honestly, the speed with which they’re launching new AI products and integrating them into their ecosystem right now is astounding. They’re constantly making new stuff. Like, I just saw this Disco thing and the live translation. And yeah, well, some of you may be thinking, well, what about the AI bubble? Fair, but it just doesn’t seem to be affecting them.
A lot of the companies I’ve talked about today have seen their stocks slow down because of all this AI bubble talk, but Google has been on an absolute tear. They’re nearing a $4 trillion valuation. And I mean, even Warren Buffett invested. At the end of the day, if AI is going to change the world like everybody says it will, then it’s hard to make an argument that there’s a better horse in this race than Google, and that’s if there was ever even a race to begin with. So, what do you think? Is there any AI company that can play on Google’s level? Leave a comment down below, and I’ll see you in the next one.
FAQs:
1. What recent Google AI release caused a significant shift in the industry?
The release of Gemini 3, which significantly outperformed OpenAI’s GPT-5 and triggered a “code red” response from competitors.
2. How is Google directly challenging Nvidia’s dominance?
By designing and mass-producing its own AI-specific chips (TPUs) and selling them directly to major AI companies, breaking Nvidia’s supply monopoly.
3. What is Google’s unique structural advantage in the AI “stack”?
It is the only company with a major presence in all five layers: chip design, data centers, foundational models, consumer products, and hardware.
4. Why does Google’s ownership of an ecosystem (Gmail, Docs, etc.) give it an edge over pure AI companies?
It allows Google to seamlessly integrate AI into products that billions already use daily, making it harder for users to switch to a competitor’s standalone service.
5. What legal event inadvertently cleared the path for Google’s aggressive AI pivot?
A U.S. antitrust case ended without breaking up Google, partly because the rise of AI (like ChatGPT) was seen as new, credible competition for its core search business.
6. What is the core financial result of Google’s vertically integrated AI strategy?
It creates massive efficiency, lowers costs compared to rivals, and is driving tremendous growth, reflected in its cloud revenue and soaring stock valuation.
