By Belle Lin

Anthropic and Databricks struck a five year, $100 million pact to sell artificial intelligence tools to businesses, targeting those seeking to build their own AI agents.

Both Anthropic and Databricks are under tremendous pressure to generate revenue from AI that can justify their sky-high valuations -- especially as they go up against some of the tech industry's biggest players, including Amazon, Microsoft, Google and OpenAI.

Based on the deal structure, Anthropic and Databricks are expected to jointly generate $100 million in revenue over a five-year period, they said. Both companies' sales teams will sell each others' products, according to Anthropic.

The companies declined to share additional financial details behind the commercial and engineering partnership.

The deal brings two of the largest players in the AI and enterprise data space together, around AI agents -- a technology that has claimed the better part of the industry's attention even as its value to businesses remains uncertain.

Anthropic's flagship Claude models will be directly available to businesses that store and analyze their corporate data in Databricks' platform, allowing them to build AI agents -- the technology that can take on tasks on behalf of humans -- using their own data, according to Ali Ghodsi, a Databricks co-founder and its chief executive.

The goal of the partnership is for the companies to better sell their technologies to large businesses, executives at Anthropic and Databricks said, giving both the opportunity to squarely target the bigger revenues those customers typically bring in.

"Databricks has built up that trust with 10,000 customers," said Kate Jensen, Anthropic's head of sales and partnerships. "Anthropic is still relatively new, but continuing to grow extremely quickly."

Still, for enterprise software vendors, and the enterprises themselves, the value of AI agents is an open question. There is a lot of " hype" around the technology, Ghodsi said, but the biggest problem with agents so far is their lack of reliability, or accuracy.

When an AI agent is asked to do something for a business, it can't be right only 50% or 70% of the time, Ghodsi said, adding that the Databricks research team is focused on helping customers get their agent accuracy rates to over 95%. "That's on par with a human worker," he said.

Other software vendors, including OpenAI, Salesforce, Amazon.com and Google, similarly sell tools for businesses to use or build their own AI agents. Databricks and Anthropic say that joint customers have asked for a more seamless integration between the two technologies.

Block, the parent company of payments-platform company Square, uses Databricks combined with Anthropic's Claude model, among others, to power its general-purpose AI agent. Thousands of Block employees are using Claude through Databricks for tasks like coding, said Jackie Brosamer, Block's vice president of data and AI platform engineering.

San Francisco-based Databricks makes money by renting out analytics, AI and other cloud-based software that taps AI-ready data for companies to build their enterprise technology tools. It raised an additional $10 billion funding round in recent months, valuing it at $62 billion, the company said.

Databricks, which was founded in 2013, bought the AI startup MosaicML in 2023 for roughly $1.3 billion and is expanding the acquired company's services -- partly through its deal with Anthropic -- to get a piece of corporate AI deals.

Ghodsi said the company is eyeing an initial public offering at the earliest next year.

Anthropic recently completed a $3.5 billion funding round that values the AI startup at $61.5 billion. The San Francisco-based startup, which has emphasized its focus on the safety of its AI models, has been focused on building up its name among large businesses seeking to use AI. It said that about half of the company's sales team focuses on selling to enterprises.

Write to Belle Lin at belle.lin@wsj.com

(END) Dow Jones Newswires

March 26, 2025 09:56 ET (13:56 GMT)

Copyright (c) 2025 Dow Jones & Company, Inc.