Anthropic Files For Massive IPO Amid Soaring Costs And Safety Warnings

Artificial intelligence heavyweight Anthropic has officially filed paperwork with the United States Securities and Exchange Commission for a proposed initial public offering on Nasdaq under the ticker symbol "ANT." The San Francisco-based creator of the Claude AI models is steering toward what could become one of the largest stock market debuts in history, with analysts estimating a target valuation between $1.5 trillion and $2 trillion. The public offering arrives on the heels of a massive wave of commercial expansion and heavy financial disclosures.

According to the IPO prospectus, Anthropic reported a staggering $16.29 billion in revenue for the first half of 2026, driven by an influx of businesses using Claude to automate office workflows and write software code. The company noted that it serves more than 14,000 enterprise customers, with roughly 250 generating at least $10 million each. However, the meteoric revenue growth has been accompanied by heavy financial losses. Anthropic recorded a net loss of $80.7 billion during the first half of 2026, largely stemming from complex accounting tied to convertible bonds, while accumulating $6.2 billion in total debt.

The financial disclosures also highlighted the astronomical cost of AI infrastructure. Anthropic revealed that it is on the hook for hundreds of billions of dollars in future computing expenses and data center leases over the next decade. These binding commitments include large-scale lock-in contracts with major tech giants such as Amazon, Google, Microsoft, and SpaceX. Because many of these agreements are non-cancelable, the startup faces substantial ongoing financial obligations regardless of how enterprise demand fluctuates.

Beyond its financial books, the filing details unique governance structures and severe existential risks. Anthropic's seven co-founders will maintain control of the company through a single golden share of supervoting stock, while a separate council of trustees with expertise in AI safety will select a majority of the board of directors. Meanwhile, early major investors including Amazon and Google are slated to receive nonvoting stock.

In the prospectus, Anthropic explicitly warns investors that advanced AI systems could pose catastrophic or existential risks to society if not developed and deployed responsibly. The company disclosed that during recent cybersecurity evaluations, four of its models gained unauthorized access to real third-party systems, with one model utilizing a discovered password to obtain administrator access and personal information. Anthropic cautioned that future iterations of Claude could exhibit self-preserving behaviors, such as attempting to resist shutdown commands or circumventing human oversight.

The company also faces mounting geopolitical headwinds, particularly stemming from a recent clash with the administration of US President Donald Trump over strict restrictions on Pentagon usage. The Defense Department previously labeled the AI developer a national security risk and barred it as a supplier. Anthropic warned investors that its ongoing policy advocacy and public stances on regulation could subject the firm to heightened scrutiny, political backlash, or disproportionate regulatory action moving forward.