Claude becomes fastest-growing AI referral source, but traffic remains tiny

Claude is the breakout name in AI referrals, but the numbers still look tiny next to search. In SE Ranking’s dataset of 101,574 websites across 250 markets, Claude’s share of traffic climbed from 0.0029% in January to 0.0141% in April, a 386% increase. Even with that surge, AI platforms together accounted for only 0.33% of traffic in April, up from 0.1976% a year earlier.

The clearest signal is not scale but speed. SE Ranking said Claude posted its largest single-month jump in March 2026, rising from 0.0049% to 0.0127% of global web traffic. That lift followed a burst of attention around Anthropic, including the Pentagon dispute over Claude’s military-use restrictions and broader debate over where the system should and should not be deployed.

That context matters because the channel is still heavily concentrated in one place. Across the full study period from January 2025 through April 2026, ChatGPT generated 78.23% of AI-referred traffic. Perplexity took 9.33%, Gemini 6.85%, Copilot 3.57%, and Claude 1.40%. In other words, Claude may be growing fastest, but it is still a small piece of an ecosystem that remains far behind organic search, which SE Ranking previously pegged at 48.5% of global internet traffic in 2025, versus just 0.15% for AI platforms.

The practical value for agencies is not volume, it is timing. Claude’s rise is an early warning signal that AI referral streams are becoming measurable enough to track by platform, content type, and brand demand. A dashboard that lumps AI traffic into one bucket misses the point. Agencies now need to separate referrals from Claude, ChatGPT, Perplexity, Gemini, and Copilot, then compare them with direct branded visits and AI citations to see which client pages are most likely to benefit before competitors start optimizing for the same surface area.

The recent spike also shows how quickly external events can move these numbers. Reuters reported on February 14 that the Pentagon was considering ending its relationship with Anthropic over Claude’s military-use restrictions. CNBC reported on February 16 and 18 that the Pentagon wanted Anthropic and other AI labs to permit use for weapons development, intelligence collection, and battlefield operations. Reuters and Associated Press later reported that Donald Trump ordered federal agencies to stop using Anthropic technology on February 27. For agencies watching AI visibility, that kind of news cycle is exactly the kind of trigger that can turn a small referral source into a useful benchmark before it turns into a larger one.