Everyone's got an opinion on which AI assistant is "the best." Fewer people have the actual numbers to back it up. So let's do what engineers do: look at the data, call out what's fuzzy, and figure out what the market share, user counts, and revenue figures actually tell us about where this industry is heading.

The Leaderboard (With Caveats)

ChatGPT remains the undisputed traffic champion. OpenAI's flagship product built a massive first-mover advantage and has defended it with aggressive product iteration—voice mode, memory, GPT-4o, the works. Hundreds of millions of monthly active users is the figure OpenAI has thrown around, though it's worth noting that "monthly active user" definitions vary wildly between companies. Someone who opened the app once counts. Keep that in mind.

Google's Gemini is the sleeper giant in this race. Google has the distribution advantage that every other player would kill for—Android, Chrome, Search, Workspace. When you pre-install your AI assistant on billions of devices, the user count conversation becomes almost unfair. The question isn't whether Gemini has users; it's whether those users are choosing Gemini or just tolerating it because it's already there. That distinction matters enormously for long-term monetization.

Anthropic's Claude has carved out a genuinely interesting niche. It doesn't have the raw user volume of ChatGPT or the distribution muscle of Gemini, but it has something arguably more valuable in the near term: enterprise credibility. Claude's context window, its Constitutional AI approach to safety, and its reputation for being less prone to confident nonsense have made it a serious contender in B2B deployments. Anthropic's revenue trajectory has reportedly been steep, with billions in annualized revenue—impressive for a company that started by explicitly not trying to win the consumer popularity contest.

Then there's Perplexity, the scrappy search-focused upstart that decided the real money was in replacing Google, not ChatGPT. Perplexity's bet is that people want AI-native search with citations rather than a chatbot that sometimes hallucinates Wikipedia. It's a smaller player by most metrics—tens of millions of users rather than hundreds of millions—but its revenue per user story is compelling, and its valuation has climbed rapidly on the back of that thesis.

Revenue: Where the Real Story Lives

User counts are vanity. Revenue is sanity. And here's where things get genuinely interesting.

OpenAI is reportedly targeting several billion dollars in annual revenue, with ChatGPT Plus subscriptions ($20/month) forming the retail backbone and API access driving significant enterprise dollars. The math works until you factor in compute costs—running inference on frontier models at scale is eye-wateringly expensive, and OpenAI has been transparent about burning cash even as revenue climbs.

Google doesn't break out Gemini revenue separately, which is a useful tell. When a product is winning, companies tend to shout those numbers from the rooftops. The Gemini integration into Google One subscriptions and Workspace is real, but the standalone monetization story remains opaque.

Anthropic is reportedly generating annualized revenue in the low billions, driven heavily by API usage and enterprise contracts. For a company with a fraction of ChatGPT's consumer brand recognition, that's a strong signal that enterprise buyers are voting with their procurement budgets.

Perplexity's revenue is smaller in absolute terms but growing fast. Their Pro subscription and advertising experiments are early-stage, but the company has been vocal about building a sustainable business rather than just chasing growth metrics. We'll see if that discipline holds as competition intensifies.

Market Share: Depends How You Slice It

Web traffic metrics put ChatGPT in a commanding lead for direct consumer engagement. But if you measure by API calls powering third-party applications, the picture gets murkier—Anthropic and OpenAI are both major players there, and Google's Gemini API is increasingly competitive on price.

Enterprise market share is the most contested and least transparent segment. Microsoft's deep integration of OpenAI models into Copilot gives OpenAI enterprise distribution it didn't have to build itself. That's a meaningful structural advantage that raw user numbers don't fully capture.

The Limiting Factors Everyone's Ignoring

Here's what the press releases skip: all four of these products are still burning money to acquire and retain users. The inference cost problem hasn't been solved—it's been partially masked by hardware improvements and model efficiency gains, but frontier AI at scale remains expensive. The company that figures out how to sustainably monetize at low cost-per-inference wins this race, not the one with the most impressive benchmark scores.

Hallucination rates also remain a real-world problem that aggregate user statistics conveniently paper over. Enterprise buyers are discovering that "mostly right" isn't good enough for regulated industries, legal work, or financial analysis. The companies building reliability and auditability into their systems—not just raw capability—are positioning for the contracts that actually matter.

Regulatory risk is the wildcard. The EU AI Act, ongoing FTC scrutiny in the US, and various data residency requirements are creating compliance overhead that favors well-capitalized incumbents. Smaller players like Perplexity will feel this disproportionately as regulations mature.

What the Numbers Actually Mean

ChatGPT won the consumer mindshare battle and has a genuine lead. Gemini is playing a long game with distribution that could prove decisive if Google executes. Claude is quietly becoming the enterprise default for organizations that take AI risk seriously. Perplexity is a genuinely interesting experiment in AI-native search that's either a future unicorn or an acqui-hire target—possibly both in sequence.

The "who's winning" question is less interesting than "winning what, exactly?" Consumer MAUs, enterprise ARR, API market share, and AI search traffic are four different games being played simultaneously. Right now, no single company is dominating all four. That's what makes 2026 genuinely competitive—and genuinely unpredictable.

Place your bets accordingly. And maybe hedge them.