By Adam Clark

Software companies have been on edge about the launch of artificial-intelligence start-up Anthropic's latest AI release. But the White House might have helped slow down the model's rollout.

Anthropic's Mythos AI model threatens to be powerful enough to automate a vast range of coding and other information-technology tasks, potentially undercutting software companies' margins. In addition, the company claims it has found thousands of "high-severity vulnerabilities" in major operating systems.

The Trump administration opposes the company's plans to widen access to Mythos, based on concerns about security risks and the strains placed on Anthropic's computing power, which could affect its use by the government, The Wall Street Journal reported, citing people familiar with the matter.

Anthropic and the Trump administration didn't immediately respond to requests for comment. A White House official told the Journal that the administration is balancing innovation and security in the AI rollout.

The administration is involved in the rollout of Mythos because of the national-security risks posed by the model.

Anthropic has offered a preview of its Mythos model to dozens of companies and was hoping to expand access to around 70 more, bringing the total to about 120, according to the report. The AI company claims Mythos is too dangerous to release to the general public currently due to the potential cybersecurity threat it poses.

A slower rollout of Mythos could be a relief for software companies that have been repeatedly hit by worries that more-capable AI models will eat into their markets. Cybersecurity companies such as CrowdStrike and Palo Alto Networks, which have been given access to the model, might also benefit from more time to evaluate its capabilities. Palo Alto was a recent Barron's stock pick.

Anthropic would ideally want to be able to fully release its most powerful models ahead of a planned initial public offering, which could come as early as this year. Anthropic -- which counts Google and Amazon.com among its investors -- was last valued at $350 billion.

However, would-be investors have approached Anthropic with proposals to participate in a fund-raising round which would value it at as much as $900 billion, according to multiple reports. Such a valuation would see it outstrip rival AI company OpenAI, last valued at $852 billion.

Anthropic appears to be making headway among U.S. users against OpenAI. By tracking transactions, research firm M Science found Anthropic rose to account for 34% of all large-language model billings for the period from April 1-19, up from 25% in March, while OpenAI fell to 57% from 64% in the same comparison period.

Write to Adam Clark at adam.clark@barrons.com

This content was created by Barron's, which is operated by Dow Jones & Co. Barron's is published independently from Dow Jones Newswires and The Wall Street Journal.

(END) Dow Jones Newswires

April 30, 2026 11:52 ET (15:52 GMT)

Copyright (c) 2026 Dow Jones & Company, Inc.