In a current LinkedIn publish, Ross Hudgens, CEO of Siege Media, argued that Perplexity’s market share is shrinking. So, together with it alongside ChatGPT, Gemini, Claude, and Google’s AI search merchandise can provide entrepreneurs a distorted image of the place their visibility really issues. Ross declared, “Everybody ought to take away Perplexity from their LLM trackers at the moment.”
I feel Hudgens is asking the proper query. I’m much less satisfied that he has the proper reply.
The AI search market is clearly concentrating round a handful of main platforms. But it surely has not reached the purpose the place search engine optimisation professionals can confidently deal with the “4 main fashions” and cease trying elsewhere.
Actually, the info suggests we could also be heading towards a three-tier market fairly than a settled leaderboard.
I’ve Seen This Film Earlier than
This argument triggered a flashback for me.
In March 2002, I attended Search Engine Methods in Boston. The dot-com crash was nonetheless recent, and I used to be working for an search engine optimisation company in Higher Boston that tracked our purchasers’ rankings throughout 15 completely different serps.
Somebody on the convention argued that we should always simplify our reporting and focus on the 5 main serps of the day, Yahoo, Excite, Lycos, AltaVista, and Ask Jeeves.
I identified that Google was lacking from the record.
That turned out to be an essential omission. Google was gaining search share quickly, regardless that it was not but one of many apparent incumbents that everybody agreed belonged on the record.
That have makes me cautious about declaring any rising search market “settled.”
The query isn’t merely which platforms are the most important at the moment. It’s which platforms are large enough, rising quick sufficient, or strategically essential sufficient that ignoring them might trigger an search engine optimisation to overlook the following shift.
See additionally: AI Search Isn’t Changing Google, It’s Layering On Prime – Similarweb Information
Perplexity Actually Is Shedding Floor
There may be loads of proof supporting Hudgens’ argument.
StatCounter’s worldwide AI chatbot referral information reveals a dramatic change throughout 2026. In June, Perplexity accounted for 7.91% of AI chatbot referral share, primarily tied with Gemini at 7.94%. By August, Perplexity had fallen to 4.31%, whereas Gemini had risen to 10.9%.
The excellence between referral share and utilization issues right here.
Similarweb’s newest publicly accessible comparability tells a special story concerning the dimension of the person platforms. Its Might 2026 information put ChatGPT at 53.9% of worldwide internet visits amongst seven main AI assistants, Gemini at 27.9%, Claude at 9.2%, DeepSeek at 4.1%, Grok at 2.4%, and each Perplexity, and Copilot at 1.3%.
That makes Hudgens’ concern about weighting comprehensible.
Suppose an search engine optimisation platform provides Perplexity the identical weight as ChatGPT regardless that Perplexity represents solely a tiny fraction of the site visitors within the related market. A very robust or weak consequence on Perplexity might certainly distort the mixture rating.
However there’s a drawback with going from that commentary to “take away Perplexity.” Small market share doesn’t essentially imply small strategic significance.
See additionally: Google Gemini Sends Extra Visitors To Websites Than Perplexity: Report
The Market Is Consolidating, However Not Round One Winner
The most important change in 2026 is just not the disappearance of Perplexity. It’s the emergence of a critical second participant.
Similarweb reported that ChatGPT’s share of AI chatbot internet site visitors fell from 76.4% a yr earlier to about 52.7% in Might 2026. Gemini rose from roughly 9% to 27.3%, whereas Claude elevated from 1.6% to eight.9%.
That may be a important redistribution of consideration. But it surely doesn’t imply ChatGPT is disappearing.
OpenAI reported it had greater than 900 million weekly energetic customers in February, and stated on the finish of July that its fashions attain a couple of billion energetic customers throughout all its merchandise.
Google, in the meantime, reported in August that the Gemini app had handed 1 billion month-to-month customers.
And Claude has established a very robust place in enterprise and developer workflows. Anthropic stated in April that greater than 100,000 clients have been working Claude on Amazon Bedrock. In August, Anthropic advised traders its annualized income had handed $65 billion and that enterprise adoption continued to develop.
These aren’t three variations of the identical enterprise. They signify three completely different distribution benefits.
ChatGPT has huge direct client adoption.
Gemini has Google’s distribution throughout Search, Android, and its broader product ecosystem.
Claude has developed a considerable place in enterprise {and professional} work.
That appears much less like a winner-take-all market than an rising oligopoly.
And Then There Is Google
That is the place I feel LLM monitoring will get conceptually messy.
Ought to Google AI Overviews and AI Mode be handled as one other LLM? For search engine optimisation functions, I might argue no. They’re extra essential than that.
Google reported in June that AI Overviews have been reaching greater than 2.5 billion customers per 30 days, whereas AI Mode had surpassed 1 billion month-to-month customers.
Google additionally stated AI Mode queries had greater than doubled each quarter since launch.
Similarweb information reported by TechCrunch confirmed AI Overviews showing in 43% of Google U.S. searches by Might 2026, up from 15% a yr earlier. AI Mode visits elevated from 126 million in June 2025 to 279 million in Might 2026.
These numbers make one factor clear. In case your goal is knowing how individuals uncover info, merchandise, firms, and web sites by way of AI, Google’s AI search expertise can’t be handled as a distinct segment LLM. It’s an AI layer contained in the world’s dominant search ecosystem.
That makes AI Overviews and AI Mode a separate monitoring class.
Claude Is The Motive Utilization Share Alone Is Not Sufficient
Claude presents one other drawback with Hudgens’ strategy.
If an search engine optimisation workforce ranks platforms strictly in line with client site visitors, Claude appears to be like comparatively small. However that may very well be deceptive for B2B firms.
Anthropic has been aggressively constructing enterprise distribution. PwC stated it could deploy Claude Code and Cowork throughout its world workforce and prepare and certify 30,000 professionals. TCS introduced a plan to make Claude accessible to 50,000 workers throughout 56 nations. Cognizant has expanded its Claude partnership and says greater than 30,000 associates have accomplished Claude coaching.
Anthropic additionally says greater than 1,000 enterprise clients have been spending greater than $1 million yearly as of April 2026.
That creates an essential distinction for search engine optimisation and content material entrepreneurs. The platform that sends probably the most referral site visitors at the moment isn’t essentially the platform that deserves probably the most consideration out of your firm. A B2C retailer, a SaaS firm, and a cybersecurity vendor might have three completely different LLM monitoring portfolios.
So, has the market consolidated?
My reply is sure, however solely partially. There may be sufficient proof to cease pretending that each AI assistant deserves equal weight.
The info more and more helps a core group consisting of ChatGPT, Gemini, and Claude, with Google AI Overviews and AI Mode handled as a separate, and arguably bigger, search-discovery layer. Perplexity belongs in a special class.
It’s now not affordable to deal with Perplexity as a peer to ChatGPT or Gemini in an general market-share calculation. However I might not delete it from the tracker. I might de-weight it.
That distinction issues.
Perplexity’s personal trajectory is value watching. Similarweb’s August 2026 evaluation nonetheless described Perplexity as an energetic participant within the AI search market, together with its technique of remaining ad-free whereas pursuing subscriptions and enterprise offers. Similarweb additionally partnered with Perplexity to make its market intelligence accessible inside Perplexity Laptop.
That doesn’t show Perplexity will regain share. It does exhibit why “small” and “irrelevant” aren’t synonyms.
What search engine optimisation Professionals Ought to Monitor Now
I might set up AI visibility measurement into three tiers.
Tier one: Ought to be the platforms which have each scale and strategic significance. Monitor ChatGPT and Gemini individually, fairly than averaging them right into a generic LLM rating. Add Claude for B2B, enterprise, {and professional} audiences.
Tier two: Ought to be AI search embedded in established ecosystems. Monitor Google AI Overviews and AI Mode individually from ChatGPT, Gemini, and Claude. The objective right here isn’t merely measuring an LLM. It’s measuring how AI is altering the search journey. Microsoft Copilot can also belong on this tier for organizations with substantial Microsoft 365 publicity. Microsoft says its first-party Copilot household has greater than 150 million month-to-month energetic customers, whereas its broader AI options throughout Microsoft merchandise attain 900 million month-to-month energetic customers.
Tier three: Ought to be rising or specialised platforms. That is the place Perplexity, Grok, DeepSeek, and different platforms belong.
Don’t give them equal weighting. Don’t ignore them both. As an alternative, monitor them for uncommon visibility, referral site visitors, citations, and progress.
Don’t Let The Common Disguise The Sign
There may be one a part of Hudgens’ argument that I strongly agree with. An mixture LLM visibility rating can create a false sense of precision.
Think about {that a} model has a 40% quotation price on ChatGPT, 35% on Gemini, 30% on Claude, and 90% on Perplexity. A easy common produces 48.75%.
However that quantity doesn’t inform a marketer a lot.
If Perplexity accounts for under a small share of the viewers or referrals that matter to the enterprise, its 90% rating mustn’t have the identical affect because the ChatGPT consequence.
The higher query is just not “What’s our common LLM visibility?” It’s “The place are the individuals who matter to this enterprise really encountering our model?”
That requires connecting three datasets.
First, measure viewers publicity. Have a look at utilization, visits, and platform distribution.
Second, measure visibility. Monitor mentions, citations, linked URLs, and the prompts that produce them.
Third, measure enterprise impression. Join AI referrals to engagement, leads, gross sales, and different conversions wherever the analytics permit it.
That produces a way more helpful measurement mannequin than a easy leaderboard.
My Verdict On Perplexity
So, I feel Ross is correct about the issue. I don’t suppose he’s fairly proper concerning the prescription.
Don’t take away Perplexity out of your LLM tracker. Take away it from the place of equal weighting.
That’s an essential distinction.
If Perplexity represents 1% of your AI referrals, giving it 25% or 33% of your visibility rating is indefensible. If it represents greater than 5% of your AI referrals, as Hudgens himself suggests, the case for maintaining it turns into a lot stronger.
And if a specific vertical will get disproportionate visibility from Perplexity, deleting it might erase a helpful aggressive sign.
The identical precept applies to each new AI platform that comes alongside.
That’s the lesson I took from that convention in Boston in early 2002. The error wasn’t monitoring too many serps. The error was assuming that at the moment’s market leaders would nonetheless be tomorrow’s market leaders.
Google was simple to miss as a result of it wasn’t but one of many 5 engines everybody had agreed mattered.
20 years later, the lesson remains to be related.
Monitor the leaders. Weight them in line with actuality. However preserve watching the outsiders which might be gaining floor.
As a result of in search, the platform you cease monitoring will be the one you would like you hadn’t.
Extra Sources:
Featured Picture: Gorodenkoff/Shutterstock
