Investors eagerly await key Anthropic filing to see blockbuster IPO will stoke demand

Investors may learn more about Anthropic’s planned initial public offering (IPO) as soon as Friday. The big question is whether the official information about the AI company will attract even more buyers or make investors think twice before buying its stock, On The Money has learned.

Wall Street sources say Friday could be the day Anthropic makes its final S-1 filing, the official document that explains its planned stock sale, public. The company privately submitted a draft S-1 to securities regulators in June, and parts of it have leaked to the press. Morgan Stanley and Goldman Sachs are reportedly leading the stock sale as the main banks helping the company go public. The IPO could raise as much as $100 billion from investors, which could value the company at as much as $2 trillion.

However, Wall Street sources warn me that the exact date of the filing is uncertain. It could come early next week instead. It is also possible that it will not happen at all, given the recent ups and downs in AI-related stocks. Investors are trying to work out whether the technology can make enough money and whether company values have risen to levels that suggest a market bubble.

Still, the possibility that the IPO will not happen appears to be mostly theoretical. The banks helping with the sale believe the filing is coming soon, and the IPO is expected to follow. It is planned for sometime in mid-November.

The official S-1 will give investors a much clearer picture of Anthropic than they have had so far. For example, it will show that the company is losing money, and a great deal of it. Sources say its losses could reach $80 billion in the first half of the year.

But investors are not focused only on whether the company is making a profit. They also want to know how fast its business is growing, and that is where the details in the filing become especially interesting.

The S-1 will explain the major risks facing the company. Experienced traders are already making bets on these risks by buying and selling shares in companies that provide the equipment and services needed to run AI systems. These concerns have added to recent swings in the share prices of Crowdsource, Nvidia and even Oracle.

It is widely known that Anthropic and its rival OpenAI, two major companies building advanced AI systems, have strong positions in the market but are seeing their growth slow. Even so, Anthropic's revenue was still large, at around $16 billion in the third quarter of this year. Its revenue is growing at a healthy rate of 50%, although that rate is lower than in earlier reporting periods.

One thing investors do not yet know is how much Anthropic expects its revenue to grow in the future. That growth will be important for paying off its huge debt, which has not yet been fully included on its balance sheet. Public reports suggest the company may have at least $130 billion in these debts and other payment obligations that are not shown on the balance sheet. But investors will not know the exact amount it owes to pay for its growing operations until they can read the details in the filing, either later this week or next week.

Investors will also look for details about how Anthropic's growth compares with that of cheaper AI systems known as “open-weight” models. Anthropic builds what are called “frontier” AI models. It owns the technology and charges businesses to use it through subscriptions and payments for units of AI use called tokens.

These models are meant to be among the most advanced available. Anthropic and OpenAI spend huge amounts of money improving their systems and the information they give users. But they face growing competition from open-weight models, which people can download and change for free. As these models improve and remain cheap, they are becoming a more attractive choice for businesses.

Investors I know will study the S-1 to judge how much this competition could slow Anthropic's revenue growth. Supporters of AI often play down the threat from competing models. OpenAI appears likely to be the next company to go public.

It is also worth looking at how Goldman Sachs is using AI. As I have mentioned, Goldman is one of the main banks leading Anthropic's IPO. For now, the bank's internal systems let employees use different frontier AI models, including Anthropic's.

But people who know the firm say it is not clear whether Goldman will soon change its approach and let employees use open-weight models, which can help them work more efficiently at a lower cost.

“Goldman might soon start reducing its use of frontier models and begin promoting open-source models inside the company,” said one source who knows what the firm is considering. A Goldman press official declined to comment.

The competition between frontier and open-weight AI is just one issue investors will watch for in Anthropic's S-1. They also want the company to explain its energy use. Running AI systems requires expensive infrastructure, and the costs are expected to rise as the technology puts more pressure on the power grid. Anthropic is expected to warn investors that energy use and the availability of power are risks to its business.

Another issue is whether AI models can be controlled. Leaders across the AI industry are debating whether progress in the technology should be paused. But this debate has not stopped Anthropic from planning to raise more money to expand its business.

There will be plenty to read about in On The Money.

A spokesperson for Anthropic had no immediate comment.