How 2025 is the year Google made up for its '2024 AI mistake' that made CEO Sundar Pichai say: This is not acceptable

Google ended 2025 as the AI comeback story of the year, with its Gemini 3 model forcing OpenAI CEO Sam Altman to declare an internal "code red" in early December. The turnaround marks a dramatic reversal from the company's embarrassing stumbles just months earlier, when its AI products generated widespread ridicule and prompted tough questions about whether the search giant had permanently lost the AI race to a startup.

The architects of Google's resurgence read like a dream team: Nobel laureate Demis Hassabis leading DeepMind's technical breakthroughs, cofounder Sergey Brin returning full-time to push the pace, and Josh Woodward transforming Gemini's consumer experience. The 42-year-old Woodward took charge of the Gemini app in April while running Google Labs, orchestrating a product comeback that's rewritten the competitive landscape. Under their combined leadership, Gemini's monthly active users nearly doubled from 350 million in March to 650 million by October, closing in on ChatGPT's 800 million weekly users. By September, Gemini had knocked ChatGPT off the top spot in Apple's App Store.




The data tells a brutal story for OpenAI. According to SimilarWeb, Google Gemini generated 1.35 billion website visits in November, marking a 14.3% increase from October. Meanwhile, ChatGPT fell below the 6 billion benchmark it had touched in October, recording 5.84 billion visits and its second consecutive monthly decline. In the lucrative enterprise market, the shift was even more pronounced—OpenAI's market share dropped to 27% according to Menlo Ventures, while rival Anthropic leads at 40% and Gemini claimed 21%.

From 'eating glue' to eating OpenAI's lunch

The confident, widespread release of Gemini 3 was a long way from where Google was after ChatGPT's launch on November 30, 2022. At Salesforce's Dreamforce event in September, Pichai admitted OpenAI had beaten Google to releasing a chatbot. "We knew in a different world, we would've probably launched our chatbot maybe a few months down the line," he said. "We hadn't quite gotten it to a level where you could put it out and people would've been okay with Google putting out that product."

Those issues remained after Google's tentative debut of the first Gemini model in December 2023, when the company faced intense backlash over "woke" outputs and ahistorical images—depicting Black Vikings and diverse Nazi soldiers in an overcorrection that became a punchline across social media. Its Gemini-powered AI Overviews in Search triggered an online furor after famously telling users to eat glue and rocks, suggesting non-toxic glue as pizza topping and recommending jumping off the Golden Gate Bridge for depression.

The mistakes were more than just mistakes. They would have threaten Google's main business. Inside Google, leaders worried that careless AI results could damage trust in Search, which brings in over $74 billion in advertising revenue each quarter. In internal meetings, Pichai made his standards clear: "This is not acceptable."

The secret weapon that broke Google's bureaucracy

What changed wasn't just the technology—it was how Google deployed it. Woodward's success stemmed from an unusual talent in Google's notoriously slow-moving culture: demolishing bureaucratic barriers. Colleagues told CNBC he mastered systems like "block," where employees flag roadblocks for his team to immediately resolve. His viral Nano Banana image generator crashed Google's own infrastructure in August, generating so much traffic that the company's custom AI chips nearly "melted," according to reports from a November all-hands meeting.

Brin's return proved equally crucial. The billionaire cofounder, who'd stepped back from daily operations in 2019, came back full-time to work alongside engineers, pushing them to work longer hours and move faster. His hands-on technical contributions combined with Hassabis's breakthroughs in AI science—work that earned Hassabis a Nobel Prize for AlphaFold—created a formidable one-two punch that OpenAI couldn't match.

The numbers that made Altman nervous

The trio delivered the knockout punch in late November with Gemini 3, an AI model that outperformed OpenAI's flagship products on industry benchmarks including SWEBench and HLE. The release sent Alphabet's stock soaring and sparked high-profile defections. Salesforce CEO Marc Benioff announced on X that after using ChatGPT daily for three years, he'd switched to Gemini 3 and wasn't "going back." "The leap is insane—reasoning, speed, images, video… everything is sharper and faster," Benioff wrote.

Within days, Altman issued his "code red" memo, telling employees they faced "temporary economic headwinds" and warning "the vibes out there will be rough for a bit." OpenAI delayed its planned advertising business, e-commerce expansion, and AI agent development to pour resources into defending ChatGPT's core experience. The company rushed out GPT-5.2 earlier this month, claiming benchmark superiority once again, but the desperation was palpable.

Why OpenAI should be worried

The AI race is far from over. OpenAI still leads in weekly active users and developer mindshare. But for the first time since ChatGPT's launch, the startup that shocked Silicon Valley is the one scrambling to keep up. Altman himself acknowledged the existential threat Google poses. Speaking on the Big Technology Podcast, he admitted that if Google had focused on crushing OpenAI back in 2023 when its AI efforts were floundering, "we would have been in a really bad place. I think they would have just been able to smash us." Now, he concedes, "Google is still a huge threat—an extremely powerful company" with one of the strongest business models in tech.

The tables have turned. OpenAI's $500 billion valuation now faces a competitor with structural advantages that very well be keeping Altman up at night. While OpenAI burns through funding rounds and has committed over $1.4 trillion in infrastructure obligations over the next eight years, Google’s AI operation runs on profits from YouTube and Search. Google controls custom AI chips, nine products with over a billion users each for distribution, and Alphabet's 62% stock surge in 2025 outpaced all megacap peers—vindicating a strategy that seemed doomed just months ago.