Anthropic spends $2.75 million on security for CEO Dario Amodei

The cost of executive safety

Anthropic's recent IPO filing revealed that the company spent $2.75 million in 2025 to provide security for its cofounder and CEO, Dario Amodei. This expenditure is notably high, as it exceeds four times Amodei's base salary of $654,167. The security bill represented more than 15% of his total reported compensation for the year, which reached $17.96 million including stock and option awards.

The company did not limit these protections to Amodei alone. Anthropic spent an additional $1.29 million to provide Daniela Amodei, the company president and cofounder, with home surveillance, private vehicles, and personal safety measures. Furthermore, the firm disclosed $97,055 for private-aircraft travel for Amodei, stating that the use of private flights was a requirement of his security program.

Anthropic has maintained round-the-clock security since 2024. The company's board determined that enhanced protection for the Amodeis and their families was necessary due to increased media attention and what it described as bona fide business-related, heightened security concerns. These measures include personal protection, home surveillance, and private cars equipped with drivers.

Rising threats to AI leaders

The company noted in its filing that heightened public attention on AI has increased both physical and informational security risks. Anthropic reported experiencing threats to employee safety and various attempts to obtain confidential information through both physical and digital means. Several of these incidents became public this year, illustrating the volatility of the current environment.

In April, a man who gained access to the company's San Francisco office informed a security officer that a high-ranking executive was in danger of being killed. During the same month, another individual who applied for a job under a false identity posted online threats to harm employees' children. Police classified that specific incident as a terroristic threat, though no arrests were made at that time.

The threats have also taken more direct forms involving weapons. A Claude user claimed in a chat to have purchased a rifle and stated that Amodei was in his sights. While the man later claimed he was joking and felt embarrassed, Anthropic reported the interaction to the police and banned the account. These incidents contribute to a broader trend where corporate backlash against AI is becoming increasingly personal.

A growing industry trend

Security intelligence firm Liferaft reported that online attacks against AI leaders and data centers increased by a factor of seven between late February and May 2026. Jonathan Graff, the CEO of Liferaft, told the Journal that the speed and severity of these threats have been surprising. The danger is moving offline as well, with some individuals carrying documents advocating for the killing of AI investors and executives.

While Anthropic's spending is high, it is not the highest in the tech world. For comparison, Meta spent approximately $22.5 million on security for Mark Zuckerberg in 2025. Anthropic's spending on Amodei is significantly lower than that of other tech giants like Alphabet, which spent $8.8 million on security for Sundar Pichai.

However, Anthropic's costs are much higher than the average corporate executive. A study by ISS-Corporate found that the median reported CEO security cost for software and services companies was $401,336. Amodei's $2.75 million bill is nearly seven times that median figure. This suggests that the high cost is tied more to the specific profile of AI leadership than to standard corporate norms.

Corporate governance and risk

The rise in executive security benefits is a broader corporate trend. According to ISS-Corporate, 28.3% of S&P 500 companies reported providing security benefits to their CEOs in 2025, which is more than double the 13.4% reported in 2021. The technology sector has seen the most dramatic increase, with the prevalence of these benefits rising 73.5% between 2020 and 2024.

As AI companies become the focus of intense public debate, their leaders are becoming the faces of the technology. Meta has previously warned that public anger toward a company can easily be transferred to the individual running it. This phenomenon was highlighted by the 2024 killing of UnitedHealthcare CEO Brian Thompson, which prompted many companies to reexamine how they protect their top leaders.

Anthropic's board regularly reviews the security measures and costs provided to the Amodeis. The firm stated that while private-aircraft use can be justified as part of a security program, boards should also consider if less costly alternatives are appropriate. Experts suggest that boards should focus on whether these expenses address legitimate business risks and whether the arrangements are regularly reassessed.