The Great A.I. Race

Introduction

Being born in 2002 means I wasn’t around, or at least not paying attention, for some of the biggest moments in tech history: the dot-com boom, the rise of personal computers, or the launch of the first iPhone. But I believe we’re living through the next great tech race right now, the battle to dominate AI, specifically large language models (LLMs).

LLMs are no longer just research projects. They’re showing up in everyday tools, writing emails, coding, answering questions, and changing how we interact with technology. And while no one knows exactly where this is all going, one thing is clear, the race is on, it’s moving fast, and LLMs are not going anywhere.

This project is a visual breakdown of that race, from its early days to July 13, 2025, when I pulled the data. To do this, I used Google Trends to track interest in the five biggest players in the space (at least in my opinion):

OpenAI (ChatGPT)
Google (Gemini)
Anthropic (Claude)
Meta (LLaMA)
Microsoft (Copilot)

The visualization also includes key events and disruptions that help explain the spikes and shifts in interest over time.

The Data

As mentioned earlier, I pulled all of the data directly from Google Trends. It was the simplest and most accessible way to track and visualize the relative popularity of each model and company over time.

That brings me to an important caveat, Google Trends uses its own proprietary popularity metric, and they don’t disclose exactly how it’s calculated. What we do know is that the data is normalized, meaning values are scaled from 0 to 100 based on the highest point of interest within the selected dataset.

In practice, this means that if one model (like ChatGPT) dominates the conversation, the others (like LLaMA) might appear barely visible on the graph. That doesn’t necessarily mean no one is using or searching for LLaMA, it just means it’s less popular in comparison to the others during that time frame.

Once I gathered the data, I uploaded it into Tableau to build the visualization. I also added reference lines to highlight major events, product launches, model upgrades, company announcements, etc., that may have influenced public interest and helped shape the trends we see.

General Reaction

As you can see, OpenAI and ChatGPT are dominating this race. That’s not unexpected, most people only know about ChatGPT, or they think “AI” and assume it’s just ChatGPT. Google, with Gemini, is a clear second, which makes sense given Google’s size, brand trust, and daily reach.

The other three, Claude, LLaMA, and Copilot, appear more interchangeable in terms of public interest, which also tracks. I’ve heard Claude mentioned more in niche communities, especially among coders, while GPT and Gemini dominate general use conversations.

The one that surprises me most is Meta. Considering how much time, money, and public attention Mark Zuckerberg has poured into AI, it doesn’t seem to be paying off, at least not in terms of mainstream visibility.

Now, onto the major moments that shaped this trendline:

Google’s rebrand from Bard to Gemini in late 2023 marked a clear spike in interest for Gemini and a slight dip for ChatGPT.
That leads into one of the biggest disruptors so far, the release of DeepSeek, a model developed by a team in China. It generated major buzz by claiming lower costs, faster development, and independence from top-tier Nvidia chips (which are largely U.S. controlled). You can see that excitement reflected in the chart. But like many flash-in-the-pan tech moments, the hype faded once people realized DeepSeek didn’t quite match the quality of the top models.
Claude 3’s release triggered another temporary dip for GPT, which makes sense given the wave of positive reviews around that model.
Soon after, Gemini 2.5 launched, sparking a longer-term upward trend for Google. In my experience, Gemini 2.5 can go toe-to-toe with any model out there. I even saw this shift firsthand in college, professors were starting to recommend Gemini over GPT for certain academic tasks.
A sharp dip in ChatGPT interest around January 1, 2025 stands out on the graph, but unlike other dips, I couldn’t find a clear explanation. No major event lined up with it, and there’s no reference line here for that reason. If you happen to know what caused it, feel free to reach out to me at jhaucone@gmail.com.
Finally, we see another big drop in GPT interest following the release of Grok 4. Because the data cuts off shortly after, we can’t see the long-term impact yet, but the initial spike is clear. Elon Musk is a polarizing figure, especially this year, and his bold claims about Grok clearly drew attention. People were, at the very least, curious enough to try it.

Predictions

I don’t expect any of these models to just disappear. For the most part, I think ChatGPT will stay at or near the top. People trust it, they like it, and it’s become deeply embedded in mainstream media and daily life. OpenAI’s partnerships also speak volumes, they’ve signed defense contracts and recently teamed up with Apple to power Siri, which could be a game-changer. Honestly, I wouldn’t be surprised if Apple drops efforts to build its own LLM and just fully adopts ChatGPT instead.

That said, I believe Gemini has the best shot at overtaking OpenAI. Google has all the resources, engineering talent, data, hardware, and distribution channels, to win this race. Gemini is already a strong model, and if Google keeps improving it at the pace they are now, I could easily see it becoming the go-to LLM for most people.

As for Claude, I think it’s carving out its own niche as a developer focused tool. That seems to be how it’s already positioned, and in the future, it may be less about competing in the general LLM space and more about replacing or complementing specialized coding tools like Cursor.

LLaMA is the wildcard. Zuckerberg has put a ton of money and effort into AI, so I find it hard to believe Meta will sit this one out. But based on the trends, it’s clear that public interest just isn’t there, at least not at the same level. Right now, people seem to prefer other options, but I expect Zuckerberg to stay involved.

Finally, there’s Copilot. I don’t think it’ll ever dominate as a standalone, general-use LLM. Instead, I expect it to continue thriving in its current form, deeply embedded in Microsoft products, helping users write code, emails, or spreadsheets more efficiently. It’s more of a productivity enhancer than a headline grabber, and that’s totally fine.

Page updated

Google Sites

Report abuse