Survey: 80% of Claude Users Earn Over $100K

A sweeping new national survey of AI usage in the United States reveals a striking demographic divide among flagship AI assistants — 80% of Claude's weekly active users come from households earning more than $100,000 per year. The findings, from a joint study by Epoch AI and Ipsos, paint a detailed picture of how America's AI landscape is fragmenting along lines of income, access, and workplace context.

The data suggests that AI assistants are no longer a monolithic category. Instead, they are rapidly stratifying by price tier, entry point, and professional use case — with higher-income users gravitating toward more advanced, premium AI services like Anthropic's Claude.

Key Takeaways From the Survey

• 80% of Claude's weekly active users belong to households with annual incomes above $100,000
• AI assistants are splitting into distinct tiers based on pricing, platform access, and work scenarios
• High-income users are adopting premium AI tools faster than any other demographic group
• Flagship large language models show measurably different user profiles from one another
• The 'AI divide' may be widening rather than narrowing as tools become more sophisticated
• Free-tier products like Meta AI and Google Gemini attract broader, more diverse user bases

Claude Emerges as the 'Premium' AI Assistant

Anthropic's Claude has long positioned itself as a thoughtful, safety-conscious alternative to OpenAI's ChatGPT. But the Epoch AI–Ipsos survey data reveals something the company may not have anticipated — or perhaps quietly encouraged. Claude's user base is overwhelmingly affluent.

The $100,000 household income threshold is significant. According to the U.S. Census Bureau, the median household income in America sits around $75,000. That means Claude's core weekly user base is drawn disproportionately from the top third of American earners.

Several factors likely contribute to this skew. Claude Pro costs $20 per month, the same as ChatGPT Plus, but Claude's distribution channels differ. Unlike ChatGPT, which benefits from massive consumer brand awareness and a widely shared free tier, Claude has grown largely through word-of-mouth in professional and technical circles. Its user acquisition funnel runs through knowledge workers, developers, and executives — demographics that naturally trend higher-income.

How Other AI Assistants Compare

The survey doesn't just spotlight Claude. It maps the user demographics across all major AI assistants, revealing clear differentiation in who uses what.

ChatGPT remains the most broadly adopted AI assistant in the U.S., with the most diverse user base across income levels. OpenAI's aggressive free-tier strategy and ubiquitous brand recognition have made it the 'default' AI tool for millions. Its income distribution more closely mirrors the general U.S. population, though it still skews somewhat toward higher earners compared to non-AI-users.

Google Gemini benefits from deep integration with Google Search, Android devices, and the broader Google Workspace ecosystem. This gives it a natural entry point for a wide demographic range, including users who may not even realize they are interacting with an AI model.

Meta AI, embedded across Facebook, Instagram, WhatsApp, and Messenger, reaches the broadest socioeconomic spectrum. Its users often encounter AI passively through social media rather than actively seeking it out.

The contrast is illuminating:

• Claude: Premium-skewing, professional, high-income, intentional adoption
• ChatGPT: Broad adoption, moderate income skew, strong brand awareness
• Google Gemini: Platform-driven, passive discovery, wide demographic reach
• Meta AI: Social-media-embedded, broadest income distribution, often passive use
• Microsoft Copilot: Enterprise-driven, tied to Office 365 and workplace deployments

The AI Access Divide Is Real — and Growing

This data raises uncomfortable questions about AI equity. If the most capable AI tools are disproportionately used by the wealthy, the productivity and knowledge advantages these tools confer could accelerate existing socioeconomic inequality.

Consider the practical implications. A corporate lawyer earning $200,000 per year uses Claude to draft briefs, analyze contracts, and synthesize case law — saving hours of billable-rate work every week. Meanwhile, a community college student relying on a free-tier chatbot gets a meaningfully less capable experience, with usage caps, slower responses, and fewer advanced features.

This is not just a matter of subscription costs. The $20 monthly fee for premium AI access is trivial for high earners but represents a meaningful expense for lower-income households. More importantly, the 'discovery gap' matters enormously. Wealthier, more educated users are more likely to hear about cutting-edge AI tools through professional networks, tech media, and workplace adoption programs.

Researchers at Epoch AI note that this stratification mirrors historical patterns in technology adoption. Early internet access, smartphone adoption, and even broadband connectivity all followed similar curves — with affluent, educated users gaining advantages first and the benefits trickling down over months or years.

Why Workplace Context Matters More Than Price

One of the survey's most nuanced findings is that workplace context may be an even stronger predictor of AI tool choice than individual income. High-income Claude users aren't just choosing the tool because they can afford it. They're choosing it because their professional environments demand or reward sophisticated AI use.

Knowledge workers in finance, law, consulting, and technology are increasingly expected to leverage AI tools as part of their workflow. Many companies are now providing enterprise AI subscriptions — Anthropic's Claude for Enterprise, OpenAI's ChatGPT Enterprise, and Microsoft Copilot for Microsoft 365 — as standard workplace tools, similar to how they provide Bloomberg terminals or Salesforce licenses.

This creates a feedback loop. Companies that invest in premium AI tools attract and retain high-earning employees. Those employees become proficient with advanced AI capabilities. They then advocate for expanded AI adoption within their organizations, further concentrating premium AI usage among high-income professionals.

The result is a professional AI ecosystem that increasingly resembles the tiered structure of financial services — where retail investors use Robinhood while institutional traders use Bloomberg.

What This Means for AI Companies

For Anthropic, the data is a double-edged sword. On one hand, a wealthy user base means higher average revenue per user, stronger willingness to pay for premium features, and a lucrative enterprise sales pipeline. Claude's positioning as the 'thinking person's AI' clearly resonates with its target market.

On the other hand, a narrow demographic base limits total addressable market growth and creates vulnerability to competitors who crack the mass-market code. If OpenAI or Google successfully combines premium capability with mass accessibility, Claude could find itself boxed into a niche.

For OpenAI, the survey validates its dual-track strategy of maintaining a robust free tier while upselling power users to ChatGPT Plus and Enterprise plans. ChatGPT's demographic breadth is a strategic asset that no competitor currently matches.

For Google and Meta, the data confirms that platform distribution trumps product excellence in reaching diverse audiences. Their AI tools may not lead benchmarks, but they reach users that standalone AI apps never will.

Key strategic implications for the industry:

• Pricing tiers will become more granular as companies target specific income segments
• Enterprise distribution will increasingly determine which AI tools dominate among professionals
• Free-tier quality will be a critical battleground for capturing lower-income and younger users
• Mobile-first strategies could help bridge the access gap, especially via Android
• AI literacy programs may become a competitive differentiator and a PR necessity

Looking Ahead: Will the AI Divide Narrow or Widen?

The trajectory of AI stratification depends largely on how quickly model capabilities commoditize. If open-source models like Meta's Llama 4 and Mistral's offerings close the performance gap with proprietary leaders, the premium advantage of tools like Claude could erode — democratizing access to high-quality AI.

However, if frontier AI capabilities continue to advance faster than open-source alternatives can follow, the gap could widen. Anthropic, OpenAI, and Google are all investing billions in next-generation models. The features these investments unlock — agentic workflows, deep research, real-time reasoning — may remain paywalled for years before becoming widely accessible.

Policymakers are beginning to take notice. Discussions around AI equity and digital inclusion are gaining traction in Washington, with some advocates calling for subsidized AI access programs modeled on the FCC's Affordable Connectivity Program for broadband.

The Epoch AI–Ipsos survey is a wake-up call. AI is not just a technology story anymore — it is becoming a socioeconomic story. The tools people use, the capabilities they access, and the productivity gains they capture are increasingly determined by how much they earn. And in a country already grappling with widening inequality, that should concern everyone — from Silicon Valley boardrooms to Capitol Hill.