For the first time since GPT-3.5 launched in November 2022, ChatGPT is no longer the only conversation that matters in AI search. Across Q1 and into Q2 2026, ChatGPT's share of generative AI traffic dropped from roughly 86.7% to 64.5%. Gemini surged to 21.5%. Perplexity held steady around 5%. Microsoft Copilot, Claude, and emerging assistants split the rest.
This is not a story of ChatGPT collapsing. ChatGPT's absolute usage continues to grow. The story is reallocation, the AI search pie is now growing faster than ChatGPT's share, and brand visibility budgets allocated entirely to ChatGPT optimization are leaving meaningful coverage on the table.
The Numbers
| Platform | Q1 2025 share | Q1 2026 share | Direction |
|---|---|---|---|
| ChatGPT | 86.7% | 64.5% | Down ~22 points |
| Gemini | 4.1% | 21.5% | Up ~17 points |
| Perplexity | 2.8% | 5.2% | Up ~2.4 points |
| Microsoft Copilot | 3.1% | 4.7% | Up ~1.6 points |
| Claude (Anthropic) | 1.8% | 2.4% | Up ~0.6 points |
| Other (Grok, You, Mistral, etc.) | 1.5% | 1.7% | Stable |
Source data combines First Page Sage market-share reports, Position Digital AI SEO statistics, and Presenc AI cross-platform monitoring. See our running market share research for the raw breakdown.
What Drove the Shift
Gemini's search integration. Google moved Gemini from a separate destination to the default AI surface inside Search, Workspace, and Android. The bulk of Gemini's share gain is not users switching from ChatGPT to gemini.com, it is users who were already on Google encountering Gemini-generated answers in surfaces they already used. The April 2026 doubling of Google AI Pro storage from 2TB to 5TB at no extra cost accelerated the bundling effect.
OpenAI's pricing pressure. The April 9, 2026 launch of the Pro $100/mo tier (clearly responsive to Claude Max's $100/mo positioning) signals that OpenAI is now fighting on revenue rather than on free-tier reach. Free-tier users with heavy usage caps look elsewhere. Many land on Gemini or Perplexity.
Perplexity's stickiness with information-density users. Perplexity's share gain looks small in absolute terms but is concentrated in exactly the audience that influences B2B procurement: researchers, analysts, journalists, technical buyers. Perplexity citations show up in vendor-evaluation memos at a rate disproportionate to its traffic share.
Microsoft Copilot's enterprise wedge. Copilot's share gain comes from Microsoft 365 enterprise rollouts, not consumer adoption. Inside enterprise customers, Copilot is the default AI assistant for Word, Excel, Outlook, and Teams. Brand mentions inside enterprise copilots drive RFP outcomes that never appear in public traffic data.
What This Means for Brand Visibility Allocation
If your AI visibility budget is 80%+ ChatGPT-focused, you are over-indexed for the 2024 distribution. The 2026 allocation that matches reality looks closer to:
- ChatGPT: ~50%, the largest single surface but no longer the only one that matters
- Gemini: ~25%, growing fastest, deeply bundled, harder to optimize for
- Perplexity: ~10%, low traffic share but disproportionate B2B influence
- Claude + Copilot: ~10%, both are enterprise-heavy with high-revenue queries
- Open-weight tail (Llama, Qwen, DeepSeek, Kimi): ~5%, hard to monitor but increasingly material
The harder truth: optimization mechanics differ across these platforms. ChatGPT rewards canonical sources, structured data, and Wikipedia presence. Gemini rewards Google-property signals (Knowledge Graph, Maps, YouTube transcripts, Google Business Profile). Perplexity rewards recent, citation-rich pages. Claude rewards balanced, evidence-grounded content. Copilot rewards Microsoft-property signals (LinkedIn, GitHub, Bing). One generic GEO playbook does not produce equivalent outcomes across all five.
The Re-Allocation Checklist
- Run mention-rate audits on Gemini and Perplexity equal in depth to your ChatGPT audit. Most teams discover their Gemini visibility is 30-50% of their ChatGPT visibility on equivalent prompts.
- Audit your Google property coverage: Knowledge Graph entity, Google Business Profile completeness, YouTube channel transcripts, Maps presence. These are Gemini's primary trust signals.
- Pitch citation-worthy content to Perplexity. Original data, primary research, and named-author analysis get cited at higher rates than generic marketing content.
- Test on Microsoft Copilot specifically if your buyers are enterprise. Copilot inside Word and Outlook generates brand mentions inside customer workflows that never appear in your analytics.
- Add at least one open-weight model to your monitoring (DeepSeek V4 Flash or Kimi K2.6 for global brands). The tail is small in share but growing fast in deployments you cannot directly see.
For the platform-by-platform optimization detail, see our AI platform visibility guides.