Anthropic has become one of the biggest winners of the artificial intelligence boom, with revenue growth that few software companies can match. But according to Palantir Technologies Inc. (NASDAQ:PLTR), the explosive adoption of AI tools may be exposing a new problem for enterprises: the bill.
At a recent customer event, Palantir CEO Alex Karp argued that companies relying directly on large language model providers are increasingly finding themselves paying more while struggling to understand the value they are receiving. The comments, as reported by The Information, come as a growing number of large enterprises publicly discuss concerns about AI spending and usage costs.
AI’s New Cost Problem
For much of the past two years, the AI conversation centered on capability. Bigger models, better reasoning and more powerful coding assistants drove rapid adoption across corporate America.
Now, attention is shifting toward cost control.
According to The Information, executives at companies including Uber Technologies, Inc. (NYSE:UBER), ServiceNow, Inc. (NYSE:NOW) and Snowflake Inc. (NYSE:SNOW) have raised concerns about escalating AI expenses tied to the latest generation of tools, including Anthropic’s Claude Code.
The issue gained attention after Uber’s CTO said earlier this year that the company burned through its full-year AI budget within the first few months of 2026 as usage surged.
In response, some enterprises are reportedly turning to model-routing systems and usage limits to reduce spending without abandoning AI altogether.
Palantir’s Pitch
Palantir believes that shift plays directly into its hands.
Rather than asking customers to commit to a single AI provider, the company has spent years building software that allows organizations to deploy applications across multiple models while tracking performance and costs.
Karp’s argument is straightforward: enterprises should focus less on buying access to the most advanced model and more on understanding how AI is being used, what it costs and whether it is generating measurable value. “You’ll go to a large language model company and learn that they don’t care about you at all,” Karp said at the event.
Palantir’s commercial chief Ted Mabrey reportedly said customer concern around token-based AI spending has seen an “explosion” over the past six weeks.
The company’s Forward Deployed Engineers, or FDEs, work directly with customers to build customized AI applications while helping them monitor spending across different models from providers including Anthropic and OpenAI.
The Next AI Battle
The debate highlights what could become the next major battleground in enterprise AI.
Anthropic’s growth remains extraordinary. The company, founded in 2021, is reportedly projecting roughly $11 billion in revenue for the current quarter alone. By comparison, Palantir has projected approximately $7.7 billion in revenue for the full year, despite expecting around 70% growth.
But Palantir is betting that the long-term winners in AI may not be the companies building the models. Instead, they could be the firms helping enterprises manage those models, control costs and avoid becoming dependent on a single vendor.
If AI’s first phase was about access to powerful models, the next phase may be about making sure those models don’t break the budget. Palantir appears determined to position itself at the center of that conversation.
Image via Shutterstock
Login to comment