The US government has added Google, OpenAI, and Anthropic to a list of approved vendors to provide AI services to civilian federal agencies, according to Bloomberg.

The companies will offer their AI tools through the Multiple Award Schedule, a federal contracting platform that allows agencies to access vendors under pre-negotiated terms.

The General Services Administration, which oversees the platform, said vendors were evaluated based on security and performance.

This move follows executive orders from President Donald Trump focused on AI development, including the adjustment of environmental standards to boost the energy supply for data centers and a requirement that federal agencies use AI systems considered “free from ideological bias.”

🔗 Source: TechCrunch

🧠 Food for thought

1️⃣ Anthropic enters federal contracting from a position of enterprise strength

The federal approval comes as Anthropic has overtaken OpenAI in enterprise AI usage, capturing 32% market share compared to OpenAI’s 25% and Google’s 20%1.

This shift represents a dramatic change from two years ago when OpenAI dominated with 50% of the enterprise market1.

Anthropic’s growth has been driven particularly by coding capabilities, where it holds 42% of the market versus OpenAI’s 21%1.

The company’s Claude models have generated $3 billion in annual recurring revenue, demonstrating strong commercial viability that likely influenced the federal approval process2.

This enterprise success suggests federal agencies may be gaining access to AI tools that have already proven effective in business environments, potentially accelerating government adoption.

2️⃣ Federal AI procurement aligns with new ideological neutrality requirements

The vendor approvals directly support the Trump administration’s executive order requiring federal agencies to use AI that’s “free from ideological bias”3.

The administration has established “Unbiased AI Principles” that mandate AI systems produce truthful outputs without ideological manipulation, addressing concerns about potential biases in AI responses3.

This represents a significant policy shift from previous approaches that emphasized AI safety and risk mitigation to one focused on ideological neutrality in government AI deployment.