OpenAI, Google, and Perplexity Close to Host AI Technology for the U.S. Government

Three major artificial intelligence companies, OpenAI, Google, and Perplexity, are nearly approved to host their AI systems directly for the United States federal government. This development marks a big change in how government agencies may soon use advanced AI tools.

In the past, AI companies had to rely on large technology contractors that already passed strict government security reviews to provide their services to federal agencies. Companies like Microsoft, Palantir, and Amazon Web Services (AWS) acted as intermediaries. These partners hosted AI models on secure systems that met federal data protection rules. While this made it easier for the government to adopt AI quickly, it also meant the original developers had limited control over how their technology was used.

OpenAI, Google, and Perplexity Close to Host AI Technology for the U.S. Government

Now, that could change. OpenAI, Google, and Perplexity applied for an expedited security review through a federal program known as FedRAMP 20x. This initiative aims to speed up the approval process for cloud systems seeking government authorization. Under this program, companies undergoing review can seek clearance faster than usual by showing they meet strong security and compliance standards.

According to reports, these companies will receive approval at a “low impact” and pilot level. This means they could start offering their AI services directly to certain government agencies in controlled pilot settings. The approval would not automatically mean every agency must use them, but it would allow those agencies interested in testing or adopting these technologies to work with the companies without needing partners.

One benefit of this direct setup is greater independence for the AI companies. Instead of relying on third parties to host and manage their models, they can work directly with government customers. This may simplify upgrades, policy changes, and feature rollouts for government users. Additionally, it could help ensure that agencies have access to the latest versions of AI models and tools.

The shift toward direct access also comes at a time when the U.S. government is increasingly integrating AI into its operations. Agencies use AI for tasks ranging from everyday data analysis to complex research and decision support. In some cases, AI tools are involved in high-security work, requiring careful controls over data privacy and access.

See Also: Google Chrome Introduces Native Split-Screen and PDF Tools

The move has not been without challenges. Some companies have relied on partners to help them enter government markets. For example, Anthropic, another AI firm, has used Palantir’s systems to provide its tools to federal users. There have been reports of tension between Anthropic and the Pentagon, partly over how and where its AI models were applied. This situation highlights how relying on intermediaries can create complicated business and ethical issues.

If approved, direct hosting by OpenAI, Google, and Perplexity could signal a new era for government-AI partnerships. It would give agencies more choice in selecting and deploying AI solutions, while allowing the AI companies themselves to manage their own cloud systems for government work. As these efforts move forward, both sides will likely continue to balance security, innovation, and clear rules for using powerful AI technology.