Claude Surge, ChatGPT Collapse: The AI Platform Market Is Decentralizing

The Data Point That Changes Everything

Twelve months ago, ChatGPT held 76.4% of global generative AI traffic. Today it sits at 52.7%. That's a 23.7-point collapse in a single year, roughly 2 percentage points per month, and accelerating in the second half of the window. For comparison: the entire internet did not move this fast when Google overtook AltaVista.

But this isn't a story about ChatGPT's decline alone. It's about what filled the gap.

Gemini, Google's platform that barely registered in 2025, now commands 27.3% of AI traffic, an 18.4-point gain in twelve months. Claude, which held 1.6% twelve months ago, has surged to 8.9% with a staggering 2.9-point jump in the last month alone. The market is no longer dominated by one platform. It's fragmenting.

And for marketers? This changes everything about visibility, strategy, and measurement.

The question isn't whether ChatGPT is dying. The question is whether you're prepared to compete in a market where no single platform owns the majority anymore. Most teams aren't. That's the opportunity.

Why ChatGPT Fell So Fast

The decline wasn't linear. In the first six months, ChatGPT shed 11.2 points (from 76.4% to 65.2%). Then the real collapse: another 11.2 points in just six months (from 65.2% to 52.7%). The velocity is accelerating, not slowing down. That matters.

Three factors are driving this unprecedented shift:

"The market isn't choosing the best AI. It's choosing the AI that's good enough for my use case and doesn't require me to switch browsers or open a new subscription."

The Claude Moment We're Missing

Claude's growth is the most underreported story in AI right now, and it's the one that should scare OpenAI.

That's a 5.5x multiplier in twelve months. And it's accelerating, not decelerating.

What makes Claude's growth different from Gemini's is the trajectory. Gemini grew because Google distributed it. Claude is growing because users prefer it. This is organic competition, not distribution leverage. When a product grows faster with organic preference than a competitor with monopoly distribution, it's signaling something structural in the market.

The implication for marketers is direct: if you're optimizing for visibility in Claude today, you're 90 days ahead of your competitors. Citation moats are being built right now, and Claude is still early enough that there's no incumbent advantage yet. That window closes fast.

The Fragmentation Trap

The decentralization of the AI market creates an immediate problem: you can't win everywhere anymore.

Twelve months ago, a marketer could think: "We need to be visible in ChatGPT, and that covers 76% of the market." Today, the same strategy covers only 53%. The remaining 47% is split across Gemini (27%), Claude (9%), Perplexity (1.3%), Grok (2.8%), DeepSeek (4%), and others. That's a lot of market share to ignore.

That sounds manageable. But here's the trap: each platform has different citation logic, different training data windows, different query patterns, and completely different user intent distribution.

• •Gemini users search for quick answers and brand comparisons. They want convenience and integration with Google products.
• •Claude users are professionals solving complex technical problems. They want depth and accuracy.
• •Perplexity users want citations and fact-checking. They're research-focused.

The same content strategy doesn't work across all of them. Worse: most marketing teams don't have the tooling to track visibility across all six platforms simultaneously. They're still tracking ChatGPT as if it's the only game.

What "Intent Quality by Platform" Really Means

A critical caveat is hiding in the data: traffic share does not equal commercial value.

"Traffic share tells you where people go, it doesn't tell you what they're asking, how deep into the purchase journey they are, or which platform is driving the most commercial outcomes."

Someone asking ChatGPT "what's the best CRM?" is different from someone asking Claude "compare Salesforce data model vs HubSpot tech stack." Same category. Completely different commercial intent.

This is the measurement blindspot nobody's talking about. Most teams are optimizing for raw visibility (GEO score). They should be optimizing for intent-qualified visibility. And they have no way to measure the difference.

What Happens Next

Bottom Line

ChatGPT didn't fail. It just stopped being the only option. In a fragmented market, being good isn't enough. You have to be visible in the platforms where your customers actually ask questions.

And you're running out of time to figure out which ones those are.