Anthropic's leadership is moving to cement control before taking the company public. The AI safety startup's seven co-founders, including Dario Amodei and Daniela Amodei, are requesting shareholder approval for a dual-class share structure that would grant them 50.1% of voting power on most corporate decisions, according to TechCrunch.
This governance approach mirrors structures used by companies like Google, Meta, and Snap. Founders retain disproportionate control relative to equity ownership, letting them steer strategy without shareholder pressure to prioritize short-term profits. For Anthropic, the timing matters. The company has raised over $5 billion in funding, including major commitments from Google and Amazon. An IPO would dilute founder equity further unless voting rights stay locked in place.
The structure protects Anthropic's stated mission around AI safety. Public markets reward growth and revenue. They punish companies that spend heavily on research, red-teaming, and safety testing without immediate returns. A founder-controlled board can pursue long-term safety work even if quarterly earnings suffer. This reflects lessons from OpenAI's nonprofit-to-capped-profit conversion and the tensions between safety research and commercialization timelines.
Anthropic has positioned safety as its competitive moat. The company emphasizes constitutional AI, a method for training models to behave according to explicit principles. It publishes safety research openly. It's refused to train models for weapons systems. These choices cost money and slow product launches. Without voting control, founders risk pressure to abandon safety initiatives in favor of faster scaling.
The shareholder vote is not routine governance theater. It sets the terms for Anthropic's public life. Founders with 50.1% of votes can hire and fire board members, set executive compensation, approve mergers, and block hostile takeovers. They control strategic decisions without needing institutional investor sign-off. This is legal but controversial. Critics argue dual-class structures entrench management and weaken accountability to public shareholders.
Anthropic's risk is different. The company operates in a sector where regulation looms. The SEC, Congress, and international bodies are drafting rules around AI training, deployment, and safety. A public Anthropic might face pressure from shareholders to cut corners on safety compliance. Founder control offers protection against that pressure.
The ask also reflects competitive reality. OpenAI remains private (with Sam Altman holding leadership control). xAI operates under Elon Musk's direction. Google owns DeepMind. Microsoft backs OpenAI. In this landscape, Anthropic's independence and founder control are selling points to customers, employees, and regulators who value autonomy from larger tech platforms.
Approval is not guaranteed, though voting structures tend to pass when founders engineer them correctly. Anthropic will need support from existing shareholders, including Google and Amazon. Both benefit from a stable, safety-focused competitor to OpenAI. Both have incentives to let Anthropic remain independent rather than see it absorbed by rivals.
The real question is whether dual-class voting survives public market scrutiny long-term. Institutional investors increasingly oppose structures that limit their influence. Funds tracking ESG metrics often exclude companies with extreme governance imbalances. Anthropic's founders are betting that safety imperatives justify the control.