YipitData, a leading provider of data-driven market intelligence, released its Anthropic Pre-IPO Overview, following Anthropic’s October 26th S-1 filing. The analysis examines business and consumer adoption, customer spending concentration, token pricing, and competition across the AI market.

Drawing on multiple sources of business and consumer information, including proprietary and exclusive datasets, the analysis offers investors a view of Anthropic’s business across customer segments and distribution channels.

The findings show significant growth in Anthropic adoption in 2026, alongside signs that B2B adoption began to flatten through September. Technology and financial services customers account for the bulk of observed spend.

“AI has become part of how businesses and consumers work every day, and the market is moving fast,” said Allan Fromen, AI & Software Sector Head, YipitData. “Prices are falling, switching between models is easy, and new models keep arriving. The question for this next phase is how well Anthropic turns broad adoption into deeper, longer-lasting relationships with its customers.”

Key findings include:

  • Business adoption growth showed signs of flattening through September. Adoption of Anthropic rose significantly year-to-date before leveling off, highlighting the importance of spending growth among existing customers.
  • B2B spending remains concentrated among the largest customers. The top 1% of business customers account for nearly 50% of observed Anthropic B2B spend.
  • Business customer retention remains strong. Anthropic shows sustained retention of 87% after month 12 and 95% after month 24.
  • Token prices fell ~50% year over year. The decline reflects price cuts for Opus 4.5, introductory pricing becoming permanent for Sonnet 5, lower cache hit pricing for Fable 5.1, and higher cache hit rates.
  • China-based models are gaining usage on OpenRouter. China-based providers, including open models like DeepSeek and Z.ai, have gained significant usage share on a platform where users can readily switch between models.
  • Consumer subscriber growth is accompanied by stronger retention. The number of Claude subscribers rose rapidly in 2026, with newer subscriber cohorts retaining at higher rates than earlier cohorts.
  • Amazon Web Services (AWS) accounts for the majority of observed Anthropic cloud spend. While AWS remains the largest channel, spending through Google Cloud Platform has grown at a faster rate year to date.

“Business and consumer adoption offer different perspectives on Anthropic’s momentum,” said Matt Wegner, Senior Research Analyst, YipitData. “Business spending remains concentrated among a small group of customers, while Claude’s newer consumer subscriber cohorts are showing stronger retention. Tracking both provides a more complete picture of how growth is evolving across its customer base.”

The overview also examines new model adoption, B2B retention and competitive spending trends involving OpenAI.

Data and Methodology

The analysis combines data from 1,500 SMB businesses, 200 large enterprise companies, and millions of consumers, accounting for more than $4B in annual AI spend.

Findings reflect the customers, transactions and activity observed within each dataset. Panel spending shares should not be interpreted as companywide revenue shares.

For the full analysis, charts, and detailed methodology, visit: https://www.yipitdata.com/resources/yipitdata-anthropic-pre-ipo-overview.

About YipitData

YipitData is a leading provider of data-driven market intelligence, trusted by more than 650 customers, including many of the world’s largest institutional investors and Fortune 100 companies. YipitData transforms trillions of real-time consumer and business spend data points into timely, granular insights on company performance and consumer behavior. Backed by over $500M in investments from Carlyle (NASDAQ: CG) and Norwest Venture Partners, YipitData’s actionable intelligence spans sectors including software, AI, e-commerce, retail, travel, and payments. For more information, visit www.yipitdata.com.

Media gallery

About The Author