Growth across products
As of 30 September, Anthropic reported more than 11,900 customers spending at least $100,000 a year, including more than 1,900 spending over $1 million and about 250 spending $10 million or more. More than half of customers in the $100,000-plus group used both the Claude platform and a first-party app, compared with about 30% at the end of the previous year.
The filing reported net revenue retention of about 900% at the end of the third quarter, which Anthropic attributed to customers expanding their use of Claude. It also said revenue in the first half of 2026 was $16.3 billion, up from $1.2 billion in the same period a year earlier.
A connected product strategy
Anthropic presents its offer as a combination of models, a developer platform and applications, rather than a single-purpose product. It argues that advances across these layers can support broader adoption. A customer might begin with the API, add Claude Code for engineers and then introduce Cowork to other staff.
The filing points to higher spending among users of Claude Code: in the third quarter, the highest-spending 5% of Claude Enterprise customers using the coding tool spent about 20 times as much per month as the average enterprise user who did not use it. Anthropic also described Applied AI teams that help customers integrate Claude into existing technology stacks, and a venture with Blackstone and other private equity firms to place engineers in businesses to build bespoke tools.
Anthropic sells through AWS, Google Cloud and Microsoft Azure as well as directly. Under its stated arrangement, it retains the customer contract and pays the cloud provider a fee, which it says helps it avoid channel conflict and set pricing.
Relevance for Southeast Asian enterprises
For organisations in banking, telecommunications, retail and the public sector, the filing offers a view of how AI suppliers may seek to expand from a single technical use into multiple teams and workflows. Its figures suggest that broader use can coincide with higher customer spending, though they describe Anthropic’s own customer base and do not establish what other organisations will achieve.
The approach also raises practical questions for buyers: how an AI service fits existing systems, how usage and costs may change as more teams adopt it, and what happens when a provider relies on external cloud platforms. The filing does not provide Southeast Asia-specific adoption figures or outcomes.
Risks to monitor
Anthropic warns that its cloud partners may prioritise their own needs or companies they have invested in, potentially making it harder for Anthropic to secure computing capacity. It also flags customer concentration: reduced spending by, or the loss of, a major customer could harm its business if not offset elsewhere.
For enterprise buyers, the next steps are to assess expected usage and costs across teams, clarify contractual and infrastructure dependencies, and consider how to manage reliance on a single supplier. Anthropic also says it intends to focus on developing new products rather than primarily entrenching existing ones, making changes to its product range another factor to track.
This article was compiled with AI from the source below.
Source: Anthropic’s IPO filing reveals its strategy for keeping enterprise customers hooked (TechCrunch, 2026-10-26)
