With investors likely to get more details of Anthropic’s pending initial public offering as soon as Friday, the question is whether the official disclosures for the AI giant will actually stoke demand – or cause investors to think twice before tapping into stock, On The Money has learned.
Friday could be the day when the AI darling’s S-1 or official offering statement is officially made public in its final form, Wall Street sources say. The company confidentially filed its draft S-1 with securities regulators in June, parts of which have leaked to the press. Morgan Stanley and Goldman Sachs are reportedly serving as the so-called lead underwriters of a stock sale that could raise as much as $100 billion from investors, valuing the company at as much as $2 trillion.
Wall Street sources caution me on the exact timing of the filing – it could come early next week. It could also very possibly not come at all given the current volatility of stocks tied to AI as investors gauge the profitability of the technology or whether current valuations are at bubble highs.
A no-go, however, appears to fall more in the theoretical possibility lane as underwriters believe the filing is coming imminently, as is the IPO. That should be set for sometime in mid-November.
The official S-1 will put more meat on the old bones of Anthropic’s story than has been disclosed so far to investors. It will show, for example, that the company is losing money – and a lot, as much as $80 billion in the first half of the year, sources say.
The lack of profitability, however, isn’t the only thing motivating investors – it’s the growth of the business, and that’s where things get interesting in reading the document.
The S-1 will explain the company’s sizable risk factors that savvy traders are already betting on through their investments in stocks of companyies that service the AI infrastructure. That adds to the recent volatility in shares of Crowdsource, Nvidia and even Oracle.
It’s pretty common knowledge that the growth of the AI “super scalers” – Anthropic and rival OpenAI — have dominant market position and slowing growth, though still large at around $16 billion for the former in the third quarter of this year. It’s increasing at a healthy 50% clip, though that’s down from the previous reporting segments.
What isn’t known is how much the company will project for future revenue growth, which is needed to pay down its massive debt load that up until now hasn’t been consolidated on its balance sheet. These off-balance sheet IOUs could be as much as $130 billion or more based on various public accounts, but again, you won’t know exactly how much the company owes to fund its ever-growing operations until you see the fine print either later this week or next.
Another factor that investors will be eyeing is how and if the documents detail Anthropic’s growth versus cheaper so-called “open-weight” AI models. Anthropic is what is known as “frontier” AI – it owns its IP and it charges businesses for their usage through subscriptions and so-called tokens.
These are supposed to be state-of-the-art models. Anthropic and OpenAI are spending boatloads to perfect their systems and the data they provide end-users. And yet, they are increasingly facing stiff competition from the open-weight variety that is free to download and modify. The improving open-weight technology and its price point make it increasingly attractive to businesses.
Investors I know will be trying to gauge from the S-1 how much this competition might eat into Anthropic’s revenue growth as AI touts downplay the competition. (OpenAI appears to be next out of the box).
Still, AI adventures at Goldman Sachs are worth considering. Goldman, as I mentioned, is a lead underwriter for Anthropic. Currently, its internal systems only offer employees access to various frontier AI models, including Anthropic.
But people close to the firm say it’s unclear if it will soon change course and begin allowing employees to tap into open-weight models as they increase their efficiency at a lower price point.
“Goldman might soon begin to throttle down the frontier models to begin pushing open source internally,” said one of my sources with knowledge of the firm’s thinking. (A Goldman press official declined to comment.)
The frontier-vs.-open-weight debate is just one thing investors will be eyeing in Anthropic’s S-1. Energy usage is something they hope the company addresses, since the costs of AI infrastructure are not cheap and are expected to grow as the technology stresses the power grid. Anthropic is expected to alert investors that energy usage and supply are risk factors.
Another one: Whether AI models can be tamed. As we all know, there’s a debate at the highest level of the AI ecosystem whether technological advancement should go on pause, though that hasn’t stopped Anthropic’s plans to find additional capital to grow.
Yes, there’s lots of good reading coming our way at On The Money.
An Anthropic press rep had no immediate comment.
