Anthropic holds more cash than almost every U.S. company

A massive financial hoard

Anthropic recently disclosed its financial standing in an S-1 filing with the Securities and Exchange Commission. The filing revealed that the nearly six-year-old artificial intelligence lab held $112.4 billion in cash plus $19.3 billion in short-term investments as of June 30. This total of $131.7 billion places the company in a unique financial position relative to the rest of the American economy.

According to data compiled by CNBC and S&P Capital IQ, this cash pile surpasses every non-bank U.S. company except for Alphabet and Berkshire Hathaway. The scale of this liquidity is particularly notable because it exceeds the cash holdings of major tech titans like Amazon, SpaceX, Meta, and Microsoft. It even surpasses the combined cash totals of Nvidia and Apple.

The company's ability to amass such a large sum is a testament to its fundraising success in the private sector. Since its founding just six years ago, Anthropic has secured an aggregate of $141 billion in private financing. This level of capital accumulation is rare for a company that has not yet gone public.

Preparing for an IPO

Anthropic is currently in the process of preparing for its initial public offering, or IPO. While the company has not shared specific details regarding the size of its offering, reports suggest it could raise more than $100 billion. If these reports hold true, Anthropic's cash hoard could potentially jump to more than $230 billion upon completion of the sale.

The filing notes that the company will not collect the entire fundraise for itself. Instead, existing investors will also be selling into the IPO, meaning the proceeds will be split among various parties. These specific details regarding the offering size and the distribution of funds are expected to be made public in about two weeks.

This upcoming debut would be historic in terms of the amount of cash a company brings to the public market. Never before has a firm gone public with a financial reserve of this magnitude. It signals a shift in how large-scale AI infrastructure is being financed and scaled by private entities before they hit the public markets.

The cost of computing power

Despite its massive cash reserves, Anthropic faces significant and growing financial obligations. The company's first-half expenses for compute and infrastructure totaled $8.95 billion. However, the filing explains that these figures do not represent the full scale of the company's future needs because of how it handles its procurement.

Anthropic typically buys its computing power 12 to 18 months in advance. Because of this lead time, the company only recognizes expenses as that capacity actually becomes available. This means the current expense figures do not capture the full weight of the contracts the company has already signed to keep its AI models running.

The company's future commitments are substantial and represent a major claim on its capital. As of June 30, Anthropic disclosed $321.4 billion in estimated future undiscounted lease obligations. Additionally, it has $265.6 billion in expected future fixed or minimum-guaranteed commitments, totaling more than half a trillion dollars in obligations.

Future obligations and growth

These massive obligations are scheduled to commence within the next few years. The terms for these agreements range from one year to as long as 20 years. These contracts are essential for ensuring that Anthropic has the necessary hardware and infrastructure to continue developing its artificial intelligence technologies at scale.

The company's financial needs are already expanding as it enters into new agreements. After June 30, the firm signed additional and updated contracts with a minimum future payment obligation of about $245 billion. The company has not yet shared the total aggregate amount of these most recent commitments.

To fulfill these obligations, Anthropic will need to continue accumulating capital. It must do this through a combination of ongoing fundraising and generating its own profits. While its current cash pile is enormous, the sheer scale of the required computing power means the company must constantly seek new ways to fund its growth.