There’s a question that’s been floating around tech YouTube for a while now, usually dressed up in alarming thumbnails: Is Google dying? The answer you’ll hear most is yes — and usually, it’s blamed on Chinese AI. DeepSeek. Kimi. Doubao. The implication is that an army of Chinese models is quietly routing traffic away from Mountain View.

That framing is half-right, half-wrong, and 100% missing the actual story. So let’s get into what the data actually says — because it’s more interesting than the clickbait version.

The Numbers Google Doesn’t Want You to Dwell On

For roughly two decades, Google’s search dominance was a settled fact of the internet — somewhere above 90% globally, year in, year out. The company was so dominant that “Google it” became a verb in seventeen languages.

That era quietly ended in 2025. Google’s global search share slipped from 92.9% in 2023 to around 89.6% by mid-2025 — the steepest multi-year decline in the company’s history. It dipped as low as 89.74% in March 2025, a level not seen since 2015. Meanwhile, AI search visits grew over 150% year-on-year by August 2025, according to Wix AI Search Lab’s monthly tracking of SimilarWeb data.

None of that sounds catastrophic in isolation. But underneath the headline numbers, something more structural is happening.

The Zero-Click Problem Is the Real Crisis

Here’s the number that actually matters: around 60% of all Google searches in 2025 ended without a single click to any external website — up from 25% just five years ago. For news-related queries specifically, that figure hit 69%, according to Similarweb’s July 2025 report tracking searches before and after AI Overviews launched. On mobile overall, the zero-click rate reaches 77%. When Google’s AI Overviews appear on a query, it jumps to 83%. In Google’s newer AI Mode, it’s 93%.

That’s not a glitch. It’s a business model shift. Google rolled out AI Overviews to all US users in May 2024. Since then, click-through rates on queries featuring an AI Overview collapsed from 15% to just 8% — a near halving. Organic CTR overall fell 61% on AI Overview queries.

Publishers have taken the brunt of this. Google search traffic to publishers dropped 33% globally in the year to November 2025. The US decline was steeper at 38%. Individual outlet damage: HubSpot down 70–80%, Business Insider down 55%, CNN down 27–38%, Forbes down roughly 50%. The New Statesman called it an “extinction-level event.” NPR used the same phrase. They’re not being dramatic — this is the primary revenue model for most independent journalism and content publishing, being dismantled in real time.

So Where Does Chinese AI Actually Fit?

This is where the YouTube videos get it muddled. Chinese AI platforms — DeepSeek, Alibaba’s Qwen, ByteDance’s Doubao, Moonshot’s Kimi — are genuinely disruptive. But they’ve disrupted the AI model market, not Google’s search business directly.

DeepSeek R1’s January 2025 launch was a genuine shock to the system. It matched OpenAI o1’s performance on key benchmarks at a fraction of the development cost, forcing a global reassessment of AI infrastructure economics. Chinese open-weight models went from roughly 1.2% of global token usage in late 2024 to nearly 30% at peak weeks by late 2025. That’s a remarkable market penetration story.

Doubao now has over 200 million monthly users. Baidu’s Ernie assistant has crossed 200 million monthly active users. These are real numbers from a real market. Chinese models have achieved over 10% market penetration in 30 countries.

But here’s the distinction the clickbait videos miss: Doubao and Kimi are primarily chat and assistant products. They’re not eating Google’s search traffic in the UK, Ireland, or the US in any measurable direct way. What they are doing is: (1) keeping Western AI labs honest on price and capability, which accelerates the overall AI adoption curve; and (2) offering a competitive AI ecosystem in markets where Western tools face geopolitical friction or cost barriers.

The Actual Competitive Map in 2026

Within AI search specifically, ChatGPT still commands about 61% of AI search visits. Gemini is closing fast at 24.8% — and notably, Gemini more than doubled its share between January 2025 and March 2026. Meanwhile, Perplexity saw a 19.7% traffic decrease in Q1 2026, and DeepSeek’s traffic declined 11.7% in the same period. The AI search market is consolidating faster than expected, not fragmenting into a dozen Chinese alternatives.

Gartner’s much-cited 2024 forecast predicted a 25% drop in traditional search volume by 2026 due to AI chatbots. That’s looking broadly accurate. More aggressive forecasts from analysts at TTMS peg 2029–2030 as the period when AI-driven search usage definitively surpasses traditional search globally.

What Changes for Content Creators and Publishers

If you’ve built a publication, a blog, a business that runs on organic Google traffic — the rules have shifted and they won’t shift back. Ranking number one on a query where AI Overviews appear now gets you a CTR of around 8%, versus 15% on the same query without one. Older SEO playbooks were written for a world where the goal was to rank. The new goal is to be cited — to become the source that AI platforms reference, summarise, and recommend.

That means E-E-A-T signals matter more than ever. Named experts, original research, primary sources, editorial transparency — these are the signals that differentiate a citable source from generic filler content. Ironically, the collapse of easy SEO traffic may end up raising the floor for content quality, even as it decimates revenue for everyone in the short term.

The other angle worth watching: branded queries have actually seen an 18% increase in clicks despite AI Overviews. People who know and trust a specific outlet are still going there directly. Which suggests that building a recognisable brand identity matters more in 2026 than it did in 2019.

My Take

The “Chinese AI is killing Google” narrative is a better YouTube thumbnail than it is an analysis. Chinese AI is genuinely impressive — DeepSeek’s cost efficiency is real, Doubao’s scale in China is real, and the pace of development from Chinese labs has been startling. But the direct mechanism by which any of that poses a near-term threat to Google search in English-speaking markets is thin.

The actual threat to Google is Google. AI Overviews is a product that solves user intent faster — but it does so by stripping the web of the traffic that funds independent publishing. That’s a slow-motion contradiction at the heart of Google’s business model: it crawls the web to power its AI, then gives answers that stop people visiting the web it crawled.

Whether that contradiction holds together long-term, whether publishers find new models, whether antitrust action reshapes the incentive structure — those are the actual questions worth tracking. Far more interesting than whatever is going on in a YouTube video with a question mark in the title.