Rising Anthropic budgets are not appearing out of thin air. Technology leaders in a recent ETR Edge Report describe where the money comes from, and the answer reaches well beyond the model wars.

Enterprise budgets for Anthropic, the developer of the Claude AI models, are rising fast. Our Edge Report on Anthropic gathered qualitative commentary from 39 technology leaders, and 35 of them plan to increase spend over the next six months. Only four plan to cut back.

The more useful question for investors is where those dollars come from. Most of the increase is new budget. The rest comes from a revealing mix of sources: OpenAI, point-solution SaaS tools, security vendors, and, in a few accounts, payroll.

Key Takeaways

• Most of the money is new. Twenty-one of 35 increasers say their added Anthropic spend is not coming at another vendor's expense.

• OpenAI is the most-named donor. Six increasers say budget is moving from OpenAI to Anthropic, and no other vendor is named as often.

• The displacement reaches past AI. Some accounts are funding Anthropic by retiring single-purpose SaaS subscriptions, reallocating security budgets, or reducing headcount.

• Cost is the pressure point. Several increasers flag rising costs or unproven ROI, and the decreasers cite pricing changes, reliability, and policy risk.

The findings come from Edge Reports, the qualitative research series from Enterprise Technology Research (ETR) that collects verbatim commentary on why technology leaders plan to change spend with a specific vendor. Because each report includes a minimum level of participation from both increasers and decreasers, the split should not be read as representative of the broader market.

Most of the Money Is New

For the majority of this panel, Anthropic spend is additive. Twenty-one increasers describe it as net new investment, expanded usage, or spend running in parallel with other AI vendors. The drivers are familiar: agentic workflows moving from pilot to production, enterprise-wide license rollouts, and coding tools spreading across engineering teams.

Coding remains the most reliable way in. Six increasers point to Claude Code, Anthropic's agentic coding tool, or its other coding tools as a driver, and several describe rollouts reaching well beyond engineering. A Department Head at a large EMEA IT/TelCo company said every desk worker now has access to Anthropic tools, "so as they get comfortable, their spend will increase."

Once inside, Anthropic is moving up the stack. A Practitioner at an EMEA Global 1000 Services/Consulting company said the plan is for Anthropic to become "the central AI layer that other systems plug into," and several other accounts describe it as their primary or core AI platform.

Is Anthropic Taking Budget From OpenAI?

When the money does come from somewhere else, OpenAI is the name that comes up most. Six increasers say they are moving budget from OpenAI to Anthropic, citing model capability, data privacy, and performance on high-stakes work. "We are going away from OpenAI and its spend is being diverted to Anthropic," said a C-level leader at a small North American company.

Performance is what keeps the budget there. An Advisor at a small EMEA IT/TelCo company said that "for as long as Anthropic has the highest performing models, we will continue to prefer using them for high-stakes tasks where we need to be able to trust the results." That condition matters. Loyalty in this market depends on staying ahead on model quality, and some accounts still run OpenAI in parallel.

SaaS and Security Budgets Are Also in Play

The more consequential signal for software investors sits outside the model layer. An Executive at a midsize North American Healthcare/Pharma company said Anthropic's models are "replacing narrower single-purpose SaaS subscriptions that previously handled tasks like document review, drafting, and process automation," and consolidating that spend into "a more flexible AI platform."

Security budgets are moving too. A C-level leader at a Fortune 1000 Services/Consulting company is adopting Claude for source code vulnerability scanning and funding it by reallocating money from other security software providers.

These are individual accounts, not a sector-wide trend, but they show general-purpose AI spend reaching into application budgets.

Is Anthropic Spend Replacing Headcount?

Some of the budget is not coming from software at all. Three increasers explicitly connect their Anthropic spend to a smaller workforce. A Senior Director at a Fortune 500 Financials/Insurance company said that, more than other vendors, "its cost is being paid by laying off staff and temporary contractors." A Practitioner at a large Education organization put it plainly: "We want to reduce human capital."

A fourth account, a Fortune 500 Services/Consulting firm rolling Claude Code out across engineering, describes the spend as net new, noting "there may be other opex reductions but not a 1:1 replacement."

Cost Is the Pressure Point

Rising spend comes with complaints about price. "Anthropic just charges more," said the same Education Practitioner, who funds both OpenAI and Anthropic. Still, one increaser is using more of Anthropic's more expensive models, a sign some buyers will pay up for capability.

ROI is the bigger risk. A C-level leader at a small Services/Consulting company described "exponential cost increases" as usage grows, and warned that the missing answer to "is it really helping us?" "may cause a drastic decrease in spend next year."

The four decreasers show how that risk turns into lost spend. One is scaling back after a shift from seat-based to consumption-based pricing. Another cited unpredictable behavior after new model releases: "The money no longer spent at Anthropic is going straight to OpenAI." The others pointed to a lack of ROI and a policy decision driven by U.S. government model restrictions.

What It Means

Anthropic's gains in this panel reach beyond the AI vendor race. The spend draws on OpenAI's share, but also on single-purpose SaaS tools, security budgets, and labor costs. That makes this Edge Report a useful lens on application software and services exposure as well as on the model providers themselves. The open question is durability: consumption pricing and unproven ROI could slow these increases at budget time if the productivity gains do not show up.

Access the Anthropic Edge Report

The Edge Report on Anthropic includes the commentary from all 39 accounts, including which vendors are losing budget to Anthropic and where decreasers are taking their spend. To trace exactly whose budgets are funding the shift, contact sales@etr.ai.