Washington doesn't want Grok: why Musk's chatbot flopped in the US government market

Grok, a chatbot positioned around "maximum honesty and no censorship," turns out to be of no use to the US government. It is a serious and unexpected blow to Elon Musk and his young company xAI.

What Reuters found: the numbers behind the failure

Reuters carried out a large-scale analysis of federal AI procurement. It reviewed more than 400 examples of how US government bodies used AI services and tools in 2024. In every case examined, tenders and reports named the specific technology supplier.

The result was a shock for xAI: Grok and its parent company appeared just three times out of 400+. And in each of those three cases the tasks were the most primitive ones: drafting official documents and managing social media accounts. No data analysis, no automation of critical processes, no integration with security systems. Just emails and posts.

Three out of 400+ means less than one percent of the market.

By contrast, OpenAI has managed to embed itself in US government infrastructure far more successfully. ChatGPT is used in IT processes at the FBI, the Department of Defense, the National Security Agency (NSA) and the State Department. Anthropic's Claude also shows up in government contract registries. But Grok? Musk pitched it as a radically "honest" chatbot, free of any censorship. Apparently that kind of honesty is not what the US government looks for in its IT systems.

Why Grok failed: five reasons

Several objective factors explain the failure in the government market:

• Newness and no contract history — xAI only recently entered the market and has no multi-year contracts, which government agencies usually renew and extend automatically year after year
• No security vetting — OpenAI and Anthropic have already gone through tough government procurement processes, independent security audits, and reviews by the Pentagon and intelligence agencies; for Grok, all of that still lies ahead
• Functionality and integration — real government work requires fitting new technology into legacy systems (COBOL applications, databases from the 1980s and 1990s, closed networks); Grok shows no readiness for that
• Unknown pricing — there is no public data on whether xAI offers favorable terms for long-term government contracts with guaranteed volumes and special licensing
• Macro-political context — Musk is known for openly criticizing government bureaucracy, "government interference" and "excessive regulation"; that may make his company a doubtful partner for conservative intelligence agencies and military departments

A blow to SpaceX and its IPO plans

The result is especially painful for Musk because Grok is pitched as one of the main growth assets ahead of the expected SpaceX IPO. Musk himself has publicly said the company will list at a valuation above $150 billion, which would make it one of the largest IPOs in modern history. A significant part of that valuation was supposed to come from Grok's investor appeal, as a chatbot positioned as a strong OpenAI competitor in the government and corporate segments.

But if the US government market shows this much skepticism and indifference, where will Grok grow? In the corporate market it faces tough competition from entrenched players OpenAI, Anthropic and Microsoft, which hold long-term contracts with Fortune 500 companies and established integrations. In the consumer market Grok continues to lag in features, quality and presence in popular ecosystems (Chrome, iOS, Android, Windows Copilot). There is no obvious route to explosive growth.

"Grok was positioned as honest and uncensored.

But the government doesn't need honesty. It needs reliability, proven integration, a long record of fulfilling contracts and security certification," one of the analysis's authors told Reuters.

What this means for investors and the market

Musk's ambition to turn Grok into a strategic growth asset for SpaceX now looks much less convincing to investors. At the IPO they will look at this figure and ask: 3 out of 400+ government contracts, what does that prove? Zero integrations with the country's critical infrastructure. Zero long-term contracts with guaranteed revenue. Zero track record in the enterprise segment. This doesn't look like a growing business; it looks like an experimental service.

The broader takeaway for the AI market: for critical systems, government agencies and large corporations choose neither "honesty" nor ideological positions, but proven technology, reliability and established processes. Musk may be more honest than his competitors, but in B2B and government procurement that is not a competitive advantage. It is just a marketing slogan.

A niche for "honest, uncensored" AI does exist, but it is not the government, the military or big corporations. It is retail, small startups and online communities. In that niche Grok holds the right position for now, but what guarantees its growth over the next three years?