Tuesday, November 18 saw the long-awaited launch of Gemini 3, Google's new version of AI, which has generated quite a bit of hype and rumors over the past month. Apparently, the rollout is going well and user feedback is mostly positive. That gave a boost to shares of parent company Alphabet, which were already up when it was revealed that they were bought by legendary investor Warren Buffett's Berkshire Hathaway fund in the third quarter. There is an interesting twist in the frenzied AI race: Gemini is clearly challenging the previously undisputed leader ChatGPT from OpenAI.
"Googling for old people."
When OpenAI launched its instantly popular AI chatbot ChatGPT in late 2022, the sirens at Google headquarters just didn't sound. CEO Sundar Pichai declared a "code red" - the highest level of danger in military terminology - and urgently held a series of meetings to determine how to respond to the new threat, The New York Times reported at the time.
It must have been particularly frustrating for him to see his competitor succeed, given the fact that it was Google that invented AI technology - or rather, the most important part of it, the Transformer algorithm. The company has been developing AI for a long time, but large corporations have to be cautious, while startups can afford to take risks for the sake of leadership. OpenAI CEO Sam Altman took a risk and won - all competitors, including Google, Meta, Microsoft and others immediately found themselves in the role of catching up.
Google's attempts to narrow the gap have often been accompanied by setbacks. In February 2023, Google's chatbot Bard AI claimed that the James Webb telescope was the first to discover a planet outside the solar system (it wasn't). Moreover, Google posted this response on its official social network X account, which caused a flurry of criticism and ridicule, as well as in a moment a 7% drop in the company's shares, the BBC wrote.
Bard AI is no longer with us - it had to be rebranded as Gemini. However, not everything went smoothly with the new name either - a year later, Gemini was at the center of a scandal when it started generating pictures depicting German soldiers of World War II as Asian or black. Google management had to apologize.
Another year later, in the spring of 2024, Gemini came under fire from critics again, this time for advising people to stick cheese on their pizza or eat rocks for health.
As a result, articles started coming out in late 2024 with titles like "Googling for old people. And that's a problem for Google." In general, the impression was that the clumsy technology giant had become obsolete and was falling further and further behind the fastest competitors in the field of AI, especially OpenAI, which was rapidly gaining momentum.
In addition, Google was facing an antitrust case in which the U.S. Justice Department was trying to have the company split into two or even three parts, separating its Chrome browser and Android operating system.
But this year, luck seems to have turned around for Sundar Pichai and the company he heads.
Mysterious AI
For starters, the company demonstrated that it's not "for old people" and knows how to flirt with the youth by releasing a new AI graphics editor called Gemini 2.5 Image in August, but it did so in an unusual way.
The program appeared on August 1 on the popular independent AI testing site LMArena under the "pseudonym" Nano Banana. According to the Global GPT portal, the name was invented spontaneously at 2 a.m. and adopted for something that "will definitely not be associated with Google".
Testers soon took notice of the new anonymous model, which was performing well in image creation and editing, quickly taking the top spot in its category and garnering rave reviews. "In my 20-year design career, I've never seen such a smart tool. It really understands what I want," AI-Stack quoted the graphic designer as saying.
In addition, the mysterious AI model appeared on the testing lists not every time, but randomly, adding an element of intrigue, play and excitement. The community rushed to investigate and network "detectives" came to the conclusion that a small company could not have created such a perfect product. On August 19, Logan Kilpatrick, head of Google AI Studio, further teased the public by publishing a post on the X network consisting of a single emoji - a banana.
And finally, ten days later, Google officially recognized that the mysterious AI is its new development Gemini 2.5 Image. By this time, all the world's media had managed to write about the novelty, notes AI-Stack, which created a detailed timeline of events.
The strategy has been a success - as Josh Woodward, vice president of Google Labs, revealed in late October, the number of Gemini users grew 44% in three months - plus 200 million to 650 million per month, largely due to the popularity of the Nano Banana among 18-34 year olds.
In September, the antitrust case was resolved in the most favorable way for Google - Judge Amit Mehta decided not to split the company into parts, limiting himself to much less severe antitrust measures, such as the need to share certain categories of data. Outraged critics called the decision a "light slap in the face", but investors reacted with enthusiasm and Alphabet's shares rose 9%, The Guardian wrote.
And that's when the heavy artillery came in.
If you can't win, lead
On Friday, November 14, it became known that the investment fund Berkshire Hathaway Warren Buffett acquired a stake in Alphabet worth $4.3 billion, while reducing its stake in Apple by about $11 billion. And this despite the fact that the legendary investor, nicknamed for his foresight "Oracle of Omaha", is known for his dislike of technology companies - Apple he sees primarily as a manufacturer of in-demand consumer goods, notes the Financial Times. "Buffett's vote of confidence in Alphabet underscores the turnaround in Google's fortunes in its battle to dominate the artificial intelligence market over the past year," the FT writes. What was the reason for Berkshire's decision to buy Alphabet shares, investors did not explain. It is known that Buffett tries to choose for investment companies surrounded by the so-called "economic moat", a strong competitive advantage that provides their leadership and allows to generate good profits in the long term. In the case of Google, this "moat" could be considered the company's undisputed dominance in the Internet search and advertising market. There was a risk that it would be destroyed by an antitrust lawsuit by the US Department of Justice, but just in the third quarter it became clear that this would not happen.
Also, fears that the AI boom will undermine Google's traditional advertising and search business model have yet to materialize - the company reported strong third-quarter results in October, with record revenue exceeding $100 billion for the first time.
In addition, apparently guided by the well-known principle "if you can't win, lead", Google is persistently rushing to become the AI market leader. Yes, according to Business Insider, ChatGPT has about 800 million active users per week versus Gemini's 650 million per month. But Google is diligently closing the gap, and the release of the new Gemini 3 should help that.
As with the Nano Banana, Google executives have been fueling the intrigue for some time, dropping hints here and there about the imminent release of the new model. Sundar Pichai, for example, published a post on X on November 15, in which he showed bets on the Gemini 3 release date on the Polymarket betting platform and put two "thoughtful" emojis.
And Nobel laureate Demis Hassabis of Google DeepMind wrote in the early morning hours of November 18 "it's almost 3, my favorite time" to which Elon Musk soon replied "looking forward to trying it out", clearly taking the post as a hint of Gemini 3's imminent release.
That's pretty much how it turned out, with Google's long-awaited model released on November 18. The next day on Wednesday, Nov. 19, the company's stock jumped 5%, CNBC notes , and initial reviews are very positive. "Congrats to Google on the Gemini 3, looks like a great model," Sam Altman wrote generously in X.
LMArena now ranks the Gemini 3 Pro at #1 overall among 269 models.
Along with the new AI, the company also unveiled the Google Antigravity development platform, on which, as promised, enhanced AI agents "transform from AI assistant to active partner. Agents can now autonomously plan and execute complex end-to-end programmatic tasks on your behalf, while validating their own code," Google promises.
At the close of trading on Friday, Nov. 14, Alphabet shares were up 45%, outperforming even artificial intelligence chip maker Nvidia and becoming the best-performing stock among large tech companies in 2025, the FT wrote. After the release of the new Gemini, Alphabet's growth since the beginning of the year has already amounted to more than 50%.
Looking at how OpenAI is battling to become a vertically integrated AI company, it's hard not to notice that Google already is one. In addition to its advanced AI model, it has its own Google Cloud facilities, its own TPU AI processors, a huge user base, a popular browser with AI-integrated browsing, its own hardwired devices where AI can be added (Pixel smartphones). And most importantly, a steady cash flow to fund it all, something OpenAI is nowhere near. As Martin Pearce, author of The Information's evening news column, wrote: "There's a suggestion that investors betting on OpenAI might be better off just buying Google stock."
This article was AI-translated and verified by a human editor
