Anthropic just launched tools to show companies what their AI spending is buying | TechCrunch

Open original

Anthropic has released reporting tools that let enterprise managers see what their team's Claude token spending produces, what it costs, and where workflows break down. The move is aimed at keeping big customers from wondering whether the spend is worth it.

  • Anthropic's financials show an operating profit in Q2 and 900% net revenue retention, meaning customers from Q3 2025 were spending about nine times more a year later.
  • The new reports pull internal analytics and employee surveys to show output, cost, and bottlenecks per team, with recommendations, and Anthropic says the data won't be used for training by default.
  • Jellyfish data shows Claude-based tools account for nearly a third of proposed code changes, with spending up 10x this year, now about 2% of engineering cost, which is growth many CFOs didn't budget for.

Anthropic's financials, shared with investors this week, show Claude Code working as a business. The company reported an operating profit in Q2 and 900% net revenue retention, meaning spending from the same customers in Q3 2025 grew ninefold by Q3 2026. To keep those enterprise customers around, Anthropic launched a set of reporting tools that let managers see what a team produced, what it cost in tokens, and where workflows broke down, using internal analytics and employee surveys. The reports also include recommendations for improving results, and Anthropic says the data won't be used for model training by default. It is also adding pooled group budgets so teams can manage token spend more easily. The underlying problem is that many companies are spending heavily on AI without clear returns. Venture-backed firms drive much of the token spend, but most Fortune 500 companies aren't seeing major gains yet. Jellyfish, which tracks developer tooling across 276,000 customers, found that Claude Code and other Anthropic-based tools account for nearly a third of proposed code changes submitted for review. Spending on those tools has grown 10x since January and now runs about 2% of total engineering cost. Nicholas Arcolano, Jellyfish's chief scientist, said that growth is what most CFOs didn't budget for, and that the conversation has shifted from spending to optimizing. Anthropic's lead looks less like a product edge than an installed base. Jellyfish's product lead, Nikolas Albarran, said Codex and OpenAI's Astra model have generated a lot of noise on X, but Anthropic's numbers haven't slowed. His explanation: teams have already standardized workflows around Claude, and there's little reason to switch. Switching costs rise further when a lab embeds forward-deployed engineers inside a customer, and the new reporting tools look like a lighter version of that hands-on help. Albarran argued that even similar models differ in quirks, such as how they handle agent.md files, so the cost to switch is never zero.

Ask about this article