Source: Anthropic's annualized revenue grew from $1B at the end of 2024 to $1.4B in early March; Manus uses tools including Claude 3.7 Sonnet to power its agent
The Information :
Context & Ripple Effects
This report is an early marker in Anthropic’s commercial arc: later coverage described annualized revenue reaching roughly $3B by late May, attributed to business demand. It also places Manus among the agent builders turning foundation-model capabilities into an end-user product.
The Manus connection became more consequential as the company later reported crossing $100M ARR within eight months of launch, showing how quickly an application layer can scale on top of third-party model infrastructure.
First-order effects
- Anthropic’s reported revenue increase signals immediate demand for its model offerings, while giving the company a larger commercial base from which to serve business and developer customers.
- Manus’s agent relies in part on Claude 3.7 Sonnet, making Anthropic’s model availability and performance a direct input to Manus’s product experience rather than merely an ecosystem option.
Second-order effects
- Agent developers using external models must treat model choice as an operating and product-design decision: differentiation shifts toward orchestration, workflow design, and customer distribution when core reasoning is supplied by a provider.
- As agent usage grows, the economics of inference become more material for application companies such as Manus, increasing pressure to match model capability to task value rather than use the most capable model indiscriminately.
Third-order effects
- The pattern points toward a layered AI market in which foundation-model vendors capture usage revenue while agent companies package those capabilities into specialized workflows; the eventual division of value will depend on who owns customer relationships and margins.
- If commercial demand continues to concentrate around a small set of model providers, agent builders may gain faster access to improving capabilities but face greater exposure to supplier terms, model changes, and platform dependence.
The trend: AI agents are becoming a commercial distribution layer for foundation models, tying application growth increasingly to inference economics and upstream model suppliers.