The AI Industry Finds Common Ground in Its Opposition to Anthropic

Anthropic has managed to bring together people who usually disagree about almost everything in artificial intelligence. AI accelerationists, safety researchers, effective altruists, and technology critics have signed the same statement accusing the company of using safety as a public image while racing to build increasingly powerful AI systems.

Why Accelerationists and AI Safety Critics Are Suddenly Targeting Anthropic

The alliance is unusual. Its members do not share one vision of the future. Some want AI development slowed down. Others want it accelerated. Some fear machines becoming too powerful; others fear that a small group of companies will use safety rules to control the market.

Their common target is Anthropic. Their reasons are not necessarily the same.

A Public Challenge Ahead of Anthropic’s Planned IPO

The controversy centers on a document called the “MisAnthropic Statement.” It accuses Anthropic of warning about the dangers of recursively self-improving AI while continuing to develop powerful systems itself.

The statement argues that the company has used its safety reputation to attract investors, employees, and public support, while weakening its own safety commitments and lobbying for regulation that could benefit its business.

Its central accusation is serious: Anthropic presents itself as a responsible AI company while allegedly building the kind of concentrated power it says society should fear.

The timing matters. The criticism arrives shortly before the company’s planned initial public offering, when its reputation and claims about responsible development may receive additional attention from investors.

However, the statement is an expression of criticism, not proof that every allegation is true.

An Unlikely Alliance With Very Different Goals

The signatories include people from across the AI debate. One striking example is Émile Torres, a prominent critic of accelerationism, signing alongside Guillaume Verdon, also known as Beff Jezos, who supports rapid technological development.

That pairing illustrates the unusual nature of the coalition.

Gary Marcus, an AI researcher skeptical of claims about large language models reaching superintelligence, also signed. Researchers and academics associated with institutions including Harvard, MIT, and UCLA joined the criticism.

But agreement was not universal.

Nathan Lambert, an advocate of open AI models and founder of Trillium Labs, said he believed Anthropic was seeking power and failing to acknowledge the risks it creates. Yet he declined to sign because he considered the statement too strong.

That distinction matters. Even people who share some concerns may disagree about the language, evidence, or political consequences of making a public accusation.

The Numbers Are Not the Main Story

The statement does not present a detailed set of financial figures, market-share estimates, or measurable evidence showing how much power Anthropic has accumulated.

Instead, its strength comes from the range of people willing to attach their names to it.

Dan Hendrycks, director of the Center for AI Safety, said several AI safety groups privately agreed with the statement but feared damaging relationships or losing future funding if they signed.

His warning points to a deeper problem. If organizations responsible for questioning powerful AI companies depend on those same companies for funding or access, their independence can become difficult to maintain.

Hendrycks previously helped secure signatures from OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei on a statement saying that reducing the risk of AI extinction should be a global priority, alongside pandemics and nuclear war.

The new criticism raises an uncomfortable question: What value do broad public commitments have if companies can later be accused of acting against their spirit?

Safety Can Become a Competitive Advantage

AI safety is not only an ethical issue. It can also be a business advantage.

Companies that convince customers, employees, and investors that their systems are safer may gain trust. That trust can help attract enterprise customers, recruit talent, and build a stronger public reputation.

But the same reputation can create a conflict. If safety rules become expensive or difficult to meet, established companies may be better positioned to absorb the cost than smaller competitors.

This is the concern behind the accusation of regulatory capture: a company may support rules that appear designed to protect society but also make it harder for new competitors to enter the market.

The statement does not prove that Anthropic has successfully used regulation in this way. It does, however, identify a real question for policymakers: Are proposed safety requirements protecting the public, or are they also protecting the companies already in power?

The Signatories Have Their Own Interests

The coalition itself deserves scrutiny.

Consider Igor Babuschkin, an xAI co-founder who left the company in August 2025 to establish an investment firm focused on AI safety research. His former company remains connected to Anthropic through infrastructure arrangements. SpaceX, which owns xAI, is renting its Colossus data center to Anthropic for Claude training.

That relationship does not automatically invalidate Babuschkin’s criticism. But it shows how difficult it is to separate public positions from commercial relationships in the AI industry.

Guillaume Verdon presents another example. He is known for supporting rapid AI development and technological growth. His concern about Anthropic appears to focus less on slowing AI for safety reasons and more on the possibility that the company could use regulation to restrict competitors.

The same accusation can therefore come from people who want very different outcomes.

Some fear that Anthropic is moving too quickly. Others fear that it could become powerful enough to decide how quickly everyone else is allowed to move.

The Biggest Weakness Is the Lack of Specific Evidence

From a competition and legal perspective, the statement raises an important issue but leaves its strongest claims underdeveloped.

It accuses Anthropic of weakening safety commitments, seeking influence, and supporting self-serving regulation. Yet the text does not provide detailed evidence showing which commitments changed, which regulatory proposals were designed to benefit the company, or how those proposals would harm competitors.

Those details matter.

A company should not be judged guilty of regulatory capture simply because it supports regulation. Equally, a company should not be allowed to use public safety as a shield against legitimate questions about its market power.

A serious investigation would need to examine policy proposals, lobbying activity, changes to safety commitments, business relationships, and the actual effects of proposed rules on smaller competitors.

The coalition also has a credibility problem. Its members have different objectives, and some have commercial or ideological reasons to oppose Anthropic. That does not make their concerns false, but it means their claims should be examined individually rather than accepted as one unified judgment.

Small Businesses Should Watch the Rules, Not the Arguments

For small and medium-sized businesses, this debate may seem distant. It is not.

AI companies increasingly influence the tools businesses use to write, code, analyze information, and serve customers. If a small number of providers gain control over the technology, infrastructure, and rules governing AI, smaller businesses may have fewer choices and less negotiating power.

Business owners should avoid depending entirely on one AI provider. They should compare alternatives, keep important business information portable, and understand how changes in pricing, access, and usage rules could affect their operations.

They should also pay attention to regulation. Rules that genuinely reduce risk can benefit everyone. Rules that only large companies can afford to follow may leave smaller businesses behind.

The Real Fight Is Over Who Gets to Control AI

The MisAnthropic Statement is unlikely, by itself, to stop Anthropic’s planned IPO or significantly slow its development. Its importance lies elsewhere.

It reveals that the AI industry is no longer divided simply between people who want faster development and people who want greater safety. A new conflict is emerging over who gets to define safety, who controls the technology, and who benefits from the rules.

Anthropic’s critics may agree on the company they want to challenge, but they do not agree on what should replace its influence.

That is the central lesson. In the AI economy, safety, competition, money, and political power are becoming increasingly connected. The public should demand evidence from every company claiming to protect society—and from every critic claiming that those companies are abusing their power.

The future of AI should not be decided by corporate promises, political slogans, or a coalition of people with competing interests. It should be decided through transparent evidence, fair competition, and rules that protect the public without giving established companies control over everyone else.

Why Accelerationists and AI Safety Critics Are Suddenly Targeting Anthropic

Anthropic AI safety, MisAnthropic Statement, AI industry criticism, AI regulation, AI accelerationists, AI safety debate, Anthropic IPO, AI market power, regulatory capture, AI competition, AI policy, AI industry funding, AI governance, small business AI