Anthropic's Clash With the Trump Administration May Be Boosting Its Business, Data Suggests

The Company Everyone in Washington Wants to Fight Is Winning Anyway

There is a strange dynamic playing out in the American AI industry right now. Anthropic, the San Francisco lab behind the Claude family of models, has spent the better part of 2026 absorbing attacks from the Trump administration. It has been called a supply-chain risk. The President told every federal agency to immediately stop using its technology. Its most advanced models were forced off the market by a government directive. And yet, by almost every available business metric, Anthropic is having the best stretch of its corporate life.

In May, the company surpassed OpenAI in business spending market share for the first time, according to data from Ramp, the expense-management platform that tracks AI spending across more than 70,000 companies. It raised $65 billion at a valuation of $965 billion, topping OpenAI's valuation at the time of that round. It filed confidential paperwork for an initial public offering. And it reportedly posted its first-ever profitable quarter.

The data is raising a question that few people would have expected to ask at the start of this year: Is the fight with the Trump administration actually helping Anthropic?

How the Conflict Started and Why It Escalated

The dispute between Anthropic and the Department of Defense has been building since early 2026. At its core, the tension is about one specific question: what a company can and cannot refuse to do when a government agency comes calling with a contract.

The Department of Defense wanted Claude for broad use across military and intelligence operations, with access described as covering all lawful purposes. Anthropic drew two lines it refused to cross: it would not allow its models to be used for mass surveillance of Americans, and it would not permit fully autonomous weapons systems that operated without human oversight. These were not arbitrary positions. They reflected the company's published AI safety principles and its existing terms of service.

The Pentagon interpreted that refusal as obstruction. In February, Secretary of Defense Pete Hegseth reportedly demanded that Anthropic CEO Dario Amodei provide a signed document granting full and unrestricted access to Claude. Amodei declined. By March, the Department of Defense had labeled Anthropic a supply-chain risk, the first time that designation had been applied to an American company. A contract worth an estimated $200 million was terminated. Trump posted on social media directing every federal agency to immediately stop using the company's technology.

A federal judge later issued a preliminary injunction blocking parts of that designation. Judge Rita F. Lin found that the Department of Defense's own records showed it had labeled Anthropic a supply-chain risk because of what it called a hostile manner through the press, not for any legitimate security reason. The court found this was a clear case of illegal First Amendment retaliation.

The June Escalation: Fable 5, Mythos 5, and a New Directive

Just as Anthropic appeared to be finding its footing through the courts, the Trump administration opened a new front. On June 13, David Sacks, the venture capitalist who had served as the administration's AI czar before departing the role, published a lengthy social media post accusing Anthropic of being reckless with the release of its newest model, Fable 5, which had launched just three days earlier. Sacks did not specify which trusted partner had raised the alarm, though reporting later identified Amazon's CEO Andy Jassy as having sent a message to the White House describing security concerns.

The core complaint was that Fable 5's guardrails could be bypassed in ways that gave users access to capabilities associated with Mythos 5, Anthropic's most powerful and restricted model. Mythos had been built for a small group of trusted organizations through a program called Project Glasswing. The concern was that a jailbreak of Fable 5 could let bad actors reach that level of capability without authorization.

Anthropic disputed the characterization, calling the fears about the jailbreak overblown. But the administration moved quickly. The White House sent a directive invoking an obscure export control authority and demanded that Anthropic block all non-Americans, including its own employees, from accessing both Fable 5 and Mythos 5. Anthropic said the directive did not explain the national security concerns that prompted it. Rather than attempting to enforce nationality-based restrictions across its global workforce and user base, the company pulled both models from the market entirely.

"There is a lot of aura that comes with your model specifically being named too dangerous to use." - Ara Kharazian, Lead Economist, Ramp

What the Ramp Data Actually Shows

A Market Share Milestone in May

Ramp's spending data covers more than 70,000 businesses that use the platform to manage expenses. It is one of the more reliable windows into how corporate America allocates its AI budget, because it reflects actual dollars spent rather than self-reported survey data.

In May 2026, Anthropic's share of business AI subscriptions tracked by Ramp rose 2.5 percentage points to reach 41%. OpenAI held 39.5%, essentially flat from the prior month. That was the first time Anthropic had come out ahead in this particular measurement.

Ramp's lead economist Ara Kharazian connected this directly to the political conflict. "If anything, it'll probably boost them," he told TechCrunch. "Anthropic's best month on record, as far as business adoption, was the month that the Department of Defense labeled them a supply-chain risk. There's a lot of aura that comes with your model specifically being named too dangerous to use."

What Businesses Are Actually Buying

Ramp's data has an important nuance worth understanding. The majority of what companies spend on AI is not subscription fees but API calls, which cover token usage for tasks like software development, data analysis, and content generation. Ramp cannot always identify which specific model is being used, but in the roughly one-third of transactions where model details are visible, businesses are overwhelmingly spending on various versions of Claude Opus, especially the later releases.

Claude Code, Anthropic's command-line coding tool, has built a strong reputation among software development teams. That use case, which generates substantial API volume, appears to be a significant driver of the business spending numbers. The controversy around Mythos and Fable 5, the most powerful and most discussed models, did not have a direct impact on the spending data because those models had not yet been on the market long enough to show up meaningfully in the numbers.

The Consumer Side: App Store Rankings and Daily Users

The business data is not the only signal. On the consumer side, the pattern looks similar. According to data from SimilarWeb, Claude's daily active users grew by more than 140% between January and mid-2026. In the week immediately following the Pentagon's cancellation of its contract with Anthropic and its removal of Claude from federal agency systems, the Claude app climbed to the top of both the Apple App Store and the Android Play Store. It had been outside the top ten just a month earlier.

That kind of app store movement is difficult to explain through ordinary product improvements alone. The timing points to a controversy-driven attention spike. When people heard that the government was blocking access to a specific AI model, a meaningful number of them apparently decided they wanted to try it.

Perception, Recruitment, and the "Too Dangerous" Effect

Brand as a Byproduct of Principle

Anthropic has spent years positioning itself as the AI lab that prioritizes safety above growth. That message has always had an audience among researchers and ethicists, but it had limited traction with mainstream business buyers who tend to care more about performance and reliability than philosophical orientation.

The confrontation with the Pentagon changed the calculus in a way that no marketing campaign could have manufactured. A former xAI engineer, speaking anonymously to CNN, put it directly: "Anthropic's perception stock within the tech community went up, not down. The Pentagon issue made Anthropic look like heroes."

Alison Taylor, a professor at NYU's Stern School of Business who studies corporate strategy, described Anthropic's approach as a calculated political risk with a meaningful upside. "There's a decent chance they walk out of this looking better than anybody else," she said.

A Recruitment Advantage at a Critical Moment

The war for AI talent is expensive and relentless. Compensation packages at frontier labs run into the millions. In that environment, where financial incentives often look similar across competitors, company values and reputation become differentiating factors.

Anthropic's public stand on autonomous weapons and surveillance appears to be resonating with the kind of researchers who think carefully about where their work ends up being applied. The company already had an 80% employee retention rate and an 88% offer acceptance rate for technical roles before the conflict with the Pentagon intensified. The public dispute appears to have strengthened rather than weakened those numbers, according to Anthropic's own recruiting data. Engineers at competing organizations, including Google and OpenAI, filed a court brief in support of Anthropic's lawsuit against the administration. Google's chief scientist Jeff Dean was among the signatories. Microsoft, which has a $5 billion investment in Anthropic, also signed on.

"I've heard Anthropic talked about more than I've ever heard." - Kanjun Qiu, investor and CEO of Nvidia-backed AI startup Imbue

What Anthropic Actually Lost in This Fight

To keep the picture honest, it is worth being clear about what the conflict has cost. The government contract that was terminated was worth an estimated $200 million. A planned investment from 1789 Capital, a venture capital firm associated with Donald Trump Jr., was abandoned. Those are real losses on paper.

The June directive forcing Anthropic to pull Fable 5 and Mythos 5 from the market also carries uncertain financial weight. Ramp's data is not granular enough to show how much revenue those specific models would have generated. Mythos had been available only to a limited group since April. Fable 5 was live for three days before it was taken down. The revenue impact is likely modest in absolute terms, but the signal it sends to enterprise buyers considering long-term commitments to Anthropic's platform is harder to quantify.

The IPO pathway is also complicated. Investors in public markets tend to be wary of companies that are in active conflict with their government. That caution does not disappear just because the company's current metrics look strong. Pre-IPO linked trading instruments tied to Anthropic reportedly declined after the Fable 5 ban was announced, even as the underlying business data remained solid.

The OpenAI Contrast

The contrast with OpenAI is instructive. When the Iran conflict began in 2026, OpenAI reportedly rushed to finalize a government deal without the restrictions that Anthropic had sought. The decision positioned OpenAI as more cooperative with the administration, but it also placed the company in a different kind of light among researchers and enterprise buyers who care about how AI systems are deployed.

OpenAI still leads Anthropic in overall consumer usage by a significant margin, according to data from Sensor Tower. ChatGPT has a longer head start in the consumer market and a larger installed base. But the business spending data from Ramp shows a narrower picture, and one that shifted in Anthropic's favor precisely during the period when the government conflict was most intense.

There is a segment of enterprise buyers, particularly those in regulated industries, professional services, and technology development, for whom the question of how an AI company behaves under pressure matters. The conflict with the Trump administration has given those buyers a concrete data point about Anthropic's willingness to hold its stated positions even when the financial consequences are immediate and real.

Is Any of This Strategic, or Is It Just Coincidence?

The more interesting question is whether Anthropic's leadership understood that holding its ground could pay off commercially, or whether the business benefits are a genuine byproduct of a decision made on principle.

Dario Amodei has not framed any of this as a marketing play. His public statements have consistently focused on the substance of the dispute: that autonomous weapons without human oversight and mass surveillance of citizens represent uses of AI that Anthropic will not support, regardless of who is asking. The company's court filings make the same argument. This does not read like positioning for an audience. It reads like someone who actually believes what he is saying and is willing to pay for it.

But the effect on brand perception is real regardless of intent. When a company draws a public line and holds it under significant pressure, it creates a kind of credibility that is otherwise very hard to build. Anthropic's safety messaging had always been present in its communications. It now has a story to back it up.

Claude's daily active users grew by more than 140% between January and mid-2026, with the sharpest increases occurring in the weeks immediately following the most intense points of government conflict.

What Comes Next: IPO, Models, and the Path Forward

Anthropic sent senior representatives to Washington in mid-June to seek a resolution with the administration over the Mythos and Fable 5 ban. That conversation is ongoing. Traders in pre-IPO markets reportedly assigned a 74% probability to Fable 5 returning to availability by mid-July, suggesting that most observers expect some kind of negotiated outcome rather than a permanent shutdown.

Some White House and intelligence officials have been working in parallel to push forward a classified contract between Anthropic and the National Security Agency, which would allow Claude to be used for intelligence analysis and identifying software vulnerabilities. That track suggests there are people within the administration who want access to Anthropic's technology badly enough to find a workable arrangement despite the public conflict.

The IPO filing, made confidentially in early June, adds urgency to reaching a resolution. Public market investors will want clarity on the government relationship before committing capital. A prolonged conflict that keeps the company's most advanced models off the market is not a story that plays well in an S-1. But a story about a company that held its principles, won in court, and came out stronger on the other side might be a different matter entirely.

The Bottom Line

The data from Ramp is a single data point, and Ara Kharazian is careful to note that the picture is not complete. Ramp cannot see the full revenue impact of pulling Mythos and Fable 5. Consumer usage data from SimilarWeb and app store rankings add texture but not precision. The relationship between political controversy and long-term enterprise buying decisions is genuinely uncertain.

What the data does show, clearly and repeatedly, is that each time the conflict with the Trump administration has intensified, Anthropic's visible business metrics have not declined. In several cases they have improved. Whether that is a durable pattern or a series of attention spikes that will eventually fade back to baseline is a question the IPO process will eventually answer.

For now, the most counterintuitive story in the AI industry is that being labeled too dangerous by a government, losing a major contract, and having your best models pulled off the market may be among the most effective things that have ever happened to Anthropic's business. That is not the outcome anyone predicted at the start of this conflict. But the numbers, such as they are, point in that direction.

Related Topics: #Anthropic #Claude #AIStartups #TrumpAdministration #AIRegulation #AISafety #OpenAI #TechPolicy #AIIndustry #DarioAmodei #ClaudeAI #AIBusiness