Claude’s +855% Traffic Surge Reveals the New AI Distribution Race — While ChatGPT Scales, Anthropic Compounds

New Similarweb data exposes a fracture in the AI assistant race: percentage momentum now belongs to Claude, growth quality separates the durable from the rented, and ChatGPT’s flatness is a sign of scale — not weakness.

AI Traffic Growth — May 2025 to May 2026 (YoY, Worldwide Web)

+855%

Claude — Total Traffic YoY

+1,460%

Claude — Email Channel YoY

+450%

Gemini — Total Traffic YoY

+1.3%

ChatGPT — Total Traffic YoY

+2,940%

Grok — Paid Search YoY

-1.3%

DeepSeek — Total Traffic YoY

Source: Similarweb, ~estimates of web traffic only; not app/API/revenue

What Happened

New Similarweb data tracking year-over-year web traffic across the five leading AI assistants — from May 2025 to May 2026, worldwide — shows Anthropic’s Claude as the standout momentum story. Claude posted approximately +855% total traffic growth YoY, leading nearly every acquisition channel: Email (~+1,460%), Social organic (~+1,224%), Referrals (~+991%), Organic Search (~+920%), Gen AI referrals (~+912%), and Direct (~+824%). These are web traffic estimates — not app usage, API calls, or revenue — but the channel-level signal is structurally meaningful.

ChatGPT was essentially flat at approximately +1.3% total YoY. That figure is not a warning sign — it is the arithmetic ceiling of enormous absolute scale. ChatGPT remains by far the largest AI assistant in web traffic volume; the law of large numbers simply makes percentage growth near-impossible at that base. Gemini was the genuine #2 grower at roughly +450% total, with Display advertising up ~+969%, reflecting Google’s distribution muscle. DeepSeek reversed its early-2026 hype cycle, declining approximately -1.3% total with most channels turning negative. Grok grew ~+37% in aggregate, but that number hides a structurally fragile composition: Paid Search up ~+2,940%, while Referrals fell ~-61% and Gen AI referrals dropped ~-56%.

The raw percentages deserve one critical caveat: Claude’s triple-digit rates partly reflect a smaller starting base relative to ChatGPT’s. Percentage growth and absolute scale are two different scoreboards. What the data does reveal unambiguously is the direction and quality of growth for each player — and that divergence is where the real strategy story lives.

The key insight: Claude’s +1,460% email growth is not a vanity metric — it is evidence of a real enterprise and prosumer audience compounding through owned channels. When email outgrows social, a product has moved from viral discovery to deliberate retention. That is how durable distribution is built.

Claude — Channel Growth Breakdown (YoY ~Approx.)

Email ~+1,460%
Social (Organic) ~+1,224%
Referrals ~+991%
Organic Search ~+920%
Gen AI Referrals ~+912%
Direct ~+824%

The Structural Read

The AI assistant race is now a two-scoreboard game. The first scoreboard — absolute scale — belongs to ChatGPT, and likely will for some time. The second scoreboard — momentum and growth quality — is where the competitive map is actively being redrawn. Anthropic is winning the second scoreboard decisively, and the channel mix tells you precisely why that matters.

The most strategically significant number in the entire dataset is not Claude’s headline +855% — it is the +1,460% email growth. Email is an owned channel. It does not depend on algorithm changes, paid auctions, or platform referrals. It signals that Anthropic has built a real audience funnel: people who sought out Claude, gave permission to be communicated with, and return on their own volition. Combined with +920% organic search and +824% direct traffic, Claude’s growth profile is the pattern of a product earning distribution — not renting it.

Grok is the photographic negative of that story. A +2,940% paid search spike is not growth — it is customer acquisition spend masquerading as traction. When the same platform simultaneously loses -61% of referral traffic and -56% of Gen AI referrals, the underlying product gravity is negative. Paid search can buy clicks; it cannot buy compounding. DeepSeek demonstrates the endpoint of hype-driven growth without retention: a sharp reversal once the novelty premium evaporates. Gemini occupies a more defensible middle position, growing via Display (+969%) — expensive but brand-building — backed by Google’s structural distribution advantages in Search and Android.

Distribution Is All You Need

The AI winner won’t be the best model — it will be the best-distributed model.

When models approach capability parity, distribution becomes the only durable moat. Claude’s multi-channel organic compounding — owned audiences, search presence, referral ecosystems — is the distribution moat being built in real time. Grok’s paid-search dependency and DeepSeek’s fading hype are two cautionary proofs of the same theorem: you cannot buy distribution permanence, you have to earn it.

The Competitive Field, Ranked by Growth Quality

Claude

MOMENTUM LEADER

+855% total. Fastest across nearly every channel. Email, organic search, and direct compounding together. The enterprise/prosumer ascent is showing up in the marketing data before it will show up in revenue disclosures.

Gemini

STRUCTURAL #2

+450% total. Display-led growth is expensive but brand-building. Google’s embedded distribution in Search and Chrome gives Gemini an organic floor no startup can replicate. Durable, if not capital-efficient.

ChatGPT

SCALE INCUMBENT

+1.3% total — flat in percentage terms, dominant in absolute volume. The law of large numbers is not a weakness; it is a description of where the ceiling sits. OpenAI’s strategic risk is not growth but retention as competitors compound.

Grok

RENTING GROWTH

+37% total, but driven almost entirely by Paid Search (+2,940%). Referrals and Gen AI referrals both contracting sharply. Growing spend, not audience. This resolves badly unless the paid traffic converts to owned retention.

Scroll to Top

Discover more from FourWeekMBA

Subscribe now to keep reading and get access to the full archive.

Continue reading

FourWeekMBA