{"slug": "why-anthropic-is-helping-enterprises-spend-less-on-claude", "title": "Why Anthropic is helping enterprises spend less on Claude", "summary": "Anthropic launched new cost-control features for Claude Enterprise, including spend alerts, cost-vs-output analytics, and model defaults, to help organizations manage rising AI inference costs. The move comes as enterprises like Uber burn through AI budgets rapidly, and Anthropic aims to maintain trust despite its own revenue surge to $47 billion.", "body_md": "Anthropic just launched adult-in-the-room controls for CFOs and CIOs freaking out about AI costs.\n\nClaude Enterprise has rolled out a trio of new features to help both organizations and employees track, monitor, and control token usage, ROI, and budget utilization. Since Anthropic has established itself as the preferred AI platform among US enterprises due to its focus on safety and privacy, this is a big deal.\n\nAs a result of executives pressuring employees to adopt AI, [tokenmaxxing](https://www.thedeepview.com/articles/why-ai-s-tokenmaxxing-obsession-ran-out-of-steam) dominated the first half of 2026, but the backlash has clearly begun as AI inference costs have run out of control, and organizations are now moving to rein in token spending.\n\nUber was one of those companies that incentivized employees to use AI by launching internal leaderboards that ranked workers by how many AI tokens they used. And then the company [reportedly](https://fortune.com/2026/05/26/uber-coo-ai-spending-tokens-claude-code/) burned through its annual AI budget in the first four months of the year. I've heard other CTOs and executives report similar challenges.\n\n\"Every organization is super worried about the costs going through the roof,\" said Databricks CEO Ali Ghodsi last month at the Databricks Data + AI Summit. \"It's completely unsustainable for the organizations out there… People's hair is on fire.\"\n\nAnd of course, a lot of these organizations are using Claude. New [Ramp data from July 6](https://ramp.com/vendors/anthropic) shows that 77% of organizations that have adopted frontier LLMs use Anthropic models, up 40 percentage points in the past year.\n\nSo it's not surprising that Anthropic is working to help organizations normalize their spending. The new features in Claude Enterprise to assist with this are:\n\n**Spend alerts**: Admins are now notified when the org reaches 75% and 90% of its usage limit, so they can raise the cap before everything gets locked down and work stops. Employees receive notifications at 75% and 95% of their personal limit and can request an increase from their manager from directly within the Claude app.**Costs vs. outputs**: As part of the richer analytics in the platform, admins can now view what was produced by Claude alongside what it cost in order to do more granular ROI analysis.**Model defaults**: Admins can now set the default model that new conversations and tasks start with for Claude, Claude Code, and Claude Cowork, so employees can start with less expensive models and only upgrade to more expensive ones when needed. This can be set based on role or across the entire org.\n\nIT professionals and leaders in organizations running Claude Enterprise and Claude Teams can access these new features in the [analytics section inside settings](https://support.claude.com/en/articles/12883420-view-usage-analytics-for-team-and-enterprise-plans) in Claude. They can set the default models in the [organizational settings](https://support.claude.com/en/articles/15694740-manage-model-access-for-your-organization).\n\n## Our Deeper *View*\n\nNo AI lab has benefited more than Anthropic from out-of-control AI inference and token costs so far in 2026. Anthropic's annual revenue run rate [jumped](https://www.nytimes.com/2026/05/29/business/dealbook/anthropic-ai-openai.html) from $9 billion at the end of 2025 to $47 billion last month. So it might seem counterintuitive that the company would want to do so much to help its customers control costs and spend less on Claude tokens. However, out-of-control inference costs have turned into such a crisis, as Ghodsi noted, that Anthropic needed to do something to maintain trust with its customers. In this case, it's taking the long view, working to be a good partner, and betting on its reputation of being the safe, reliable, enterprise-friendly AI lab. Establishing that reputation could be worth far more in the long run if AI lives up to its promises.", "url": "https://wpnews.pro/news/why-anthropic-is-helping-enterprises-spend-less-on-claude", "canonical_source": "https://www.thedeepview.com/articles/why-anthropic-is-helping-enterprises-spend-less-on-claude", "published_at": "2026-07-06 22:28:35+00:00", "updated_at": "2026-07-07 01:39:05.110996+00:00", "lang": "en", "topics": ["ai-products", "ai-ethics", "ai-infrastructure", "ai-policy"], "entities": ["Anthropic", "Claude Enterprise", "Uber", "Databricks", "Ali Ghodsi", "Ramp"], "alternates": {"html": "https://wpnews.pro/news/why-anthropic-is-helping-enterprises-spend-less-on-claude", "markdown": "https://wpnews.pro/news/why-anthropic-is-helping-enterprises-spend-less-on-claude.md", "text": "https://wpnews.pro/news/why-anthropic-is-helping-enterprises-spend-less-on-claude.txt", "jsonld": "https://wpnews.pro/news/why-anthropic-is-helping-enterprises-spend-less-on-claude.jsonld"}}