U.S. adds OpenAI, Google, and Anthropic to approved AI vendor list

The U.S. government has officially designated OpenAI, Google, and Anthropic as approved artificial intelligence (AI) vendors, paving the way for easier adoption of their tools across federal agencies. U.S. 

Announced by the General Services Administration (GSA), the move is part of a broader push by the U.S. to accelerate the responsible use of AI in civilian government operations. The companies—creators of ChatGPT, Gemini, and Claude—are now available through the GSA’s Multiple Award Schedule, a contracting mechanism that offers pre-negotiated pricing and terms.

Previously, agencies had to go through time-consuming legal and procurement processes to acquire AI tools. The GSA has effectively removed much of that red tape by adding these three firms to its schedule.

“We’re not in the position of picking winners or losers here. We want the maximum number of tools to provide to all federal government employees to make them as productive as possible,” said GSA Deputy Administrator Stephen Ehikian. “There’s going to be different tools for different use cases.”

Though specific contract values remain undisclosed, the GSA is known for securing deep discounts through bulk purchasing—similar to past deals with Adobe, Salesforce, and Google.

The selected AI models underwent GSA vetting for safety, performance, and security. Officials added that more AI vendors may be approved soon, pending completion of the evaluation process.

Federal agencies prepare for widespread AI integration

For example, the Office of Personnel Management (OPM) plans to use AI to develop customer service chatbots and analyze public comments on federal regulations. OPM Director Scott Kupor noted that manually summarizing tens of thousands of citizen responses often delays the regulatory process, but said that with AI, the agency could work faster and more accurately.

Other departments like the Treasury and the Department of Commerce are exploring AI to help detect tax fraud, process patents, and support grant application reviews. Until now, many of these tools were limited to small-scale pilot projects, often siloed within national security or research-focused initiatives. The GSA’s new approval opens the door for agency-wide and cross-agency adoption.

The Pentagon has already been ahead of the curve, issuing contracts to OpenAI and Elon Musk’s xAI for military-related projects. Those deals are separate from the GSA program but signal a federal consensus: AI is now a core component of U.S. government strategy.

White House sets new standards to curb AI bias

The GSA’s announcement follows a recent push from U.S. President Donald Trump to reshape federal AI policy. Just days before the vendor list was released, Trump signed three executive orders to tighten government AI use oversight.

This rule suggests that it will be implemented at the agency level, with each department deciding what bias is and how they imagine they can test for it. Trump and other conservative leaders have regularly referred to bias as the “woke AI” problem.

The effort goes beyond simply processing new technology, said Josh Gruenbaum, Commissioner of the GSA’s Federal Acquisition Service. He called it winning the global race in artificial intelligence, the same thing the president said when he declared that “the U.S. must win.”

Still, the inclusion of OpenAI, Google, and Anthropic, companies often accused by conservatives of holding liberal biases, suggests the government focuses more on utility and performance than ideology, at least in the early stages.

The GSA noted that the agency’s vetting process included bias assessments, security checks, and performance evaluations, and added that more tools would be added as they meet the required standards.

Super Micro slashes revenue outlook by $7b

On other developments, the stock of Super Micro Computer Inc. fell 16% in after-hours trading on Tuesday. The share price dropped on news that it had drastically cut its revenue forecast for the year to June 2026. The new minimum outlook of $33 billion was much lower than the $40 billion estimate five months ago.

That surprised a few investors and analysts, given that the company was previously relatively bullish about riding the global wave of AI infrastructure demand. Super Micro was among the top players in February in artificial intelligence servers. It seemed well-positioned given its connection to chipmaker Nvidia Corp., whose GPUs are central to most AI systems.

The lowered forecast indicates that the company may have difficulty, and specifically, no longer offers any momentum amid increased competition, customer pushouts, and pricing pressures. It also raises questions about Super Micro’s ability to scale and think more in line with how the market now expects it.

Super Micro’s profit margins slide under pressure

Beyond the revenue cut, Super Micro’s profitability is also under pressure. For the current quarter ending in September, the company projected earnings per share — excluding certain items — to fall between 40 and 52 cents, well below the average analyst expectation of 59 cents.

The company also guided for revenue from $6 billion to $7 billion during the same period, again trails Wall Street’s average forecast of $6.59 billion.

According to Bloomberg Intelligence analyst Woo Jin Ho, the lower guidance points to a narrower operating margin of around 5%, compared to the 7% analysts were targeting. “Super Micro’s 1Q and 2026 outlook lowers the expectations bar,” Ho said. “However, it implies a highly competitive pricing environment in these mega-server deals, particularly from Dell.”

Super Micro reported $5.76 billion in revenue for the fiscal fourth quarter, a 7.5% year-over-year increase. But it still fell short of the $6.01 billion analysts expected. Similarly, adjusted earnings came in at 41 cents per share, missing the 44-cent estimate.