Anthropic IPO part II: They said what?
Eight days ago, I wrote here about Anthropic’s forthcoming IPO, which will likely be the second most-watched in financial market history, after SpaceX’s float in June.
What’s worrying is that since that post, a raft of Australian publications have run articles either promoting the Initial Public Offering (IPO) or advising investors on how to participate.
Unfortunately, some of these articles fail to provide anything close to comprehensive analysis, yet they confidently conclude the IPO ‘looks good’.
With many Aussies encouraged to buy into the June SpaceX IPO and those shares trading below the issue price, as well as trading 34 per cent below the high price registered two days after the float, a less than effusive look at the articles promoting the Anthropic IPO is necessary.
Misleading compounding logic & annualisation fallacies
One article highlights Anthropic’s 89 per cent quarterly compound revenue growth versus 54 per cent cost growth over seven quarters. The analyst quoted describes these as a “happy set of numbers” that proves it is a good business.
Compounding early-stage startup growth quarter over quarter and projecting it into a mature state is a classic forecasting error.
Hyper-growth in emerging, capital-saturated artificial intelligence (AI) markets is always front-loaded. To assume that it reflects long-term economic structural profitability ignores the natural decay curve of technology adoption.
Further, Anthropic reported a US$80.7 billion loss in the June half alone. Celebrating a cost growth rate below the revenue growth rate matters only if the baseline numbers are sustainable. A simple reality is that when absolute losses run in the tens of billions of US dollars, fractional cost convergence doesn’t guarantee a path to profitability. For now, all we can confirm is that it highlights an extreme rate of cash burn.
“Valuation”
In the same article, two other analysts framed the Price-to-Sales (P/S) ratio of 17x to 20x as “reasonable” or “OK,” comparing it favourably to Palantir, SpaceX, or Nvidia.
This comparison fails several basic tests.
First, Price-to-Sales is a dangerous metric for foundation model developers.
Let me explain by contrasting model developers with Software as a Service (SaaS) businesses. For a SaaS business, a customer pays, say, $1.00 to use the software, yet it might only cost $0.02-$0.08 in server costs to run that user session. The SaaS company therefore keeps $0.92 to $0.98 in gross profit (90 per cent + gross margins).
For generative AI Large Language Models (LLMs), every single user prompt triggers intensive hardware compute with Graphic Processing Units (GPUs) running inference. If a customer pays $1.00 to generate text or code, but Anthropic or OpenAI has to pay Cloud providers (Hyperscalers) $0.80 in raw hardware, compute and electricity to process those tokens, the company only keeps $0.20 in gross profit, equivalent to a 20 per cent gross margin.
The fundamental (and dramatic) difference in margin must result in a different multiple being applicable.
Meanwhile, because it looks only at top-line revenue, P/S ignores what it costs to produce that revenue. Consequently, calling a 17x–20x P/S ratio “reasonable” by comparing Anthropic to traditional high-margin tech companies (or hardware monopolies like Nvidia) fails to acknowledge the unit economics.
Further, comparing an AI model layer to Nvidia (which sells high-margin physical hardware to Hyperscalers with a supply-constrained moat) or SpaceX (which operates a near-monopoly launch service) is an apples-to-oranges comparison.
Structurally ignorant?
Meanwhile, many of the positive assessments gloss over the fundamental microeconomics of LLMs.
The first is that consumers and enterprises can switch between frontier models (OpenAI, Anthropic, Google Gemini, open-weight models like Meta’s Llama or Chinese alternatives) at zero cost.
The second is inference pricing. Competing models often reach parity every few months, and this drives aggressive price cuts for end users.
In other words, valuing software businesses without structural moats, such as network effects, high switching costs, or proprietary data lock-ins, at US$1.5 trillion – US$2 trillion ignores or forgets the reality of competitive strategies.
Fear of missing out
In another article, one investor argued that Australian investors should buy U.S. tech IPOs “whenever possible” because local indices are heavily weighted toward banks and miners. While portfolio diversification is worth encouraging, having an underallocation to global tech is not a rational reason to buy a potentially overvalued, loss-making IPO at the height of its hype cycle.
And as we saw with the failed Firmus IPO, institutional investment banks generally pass down allocation shares to retail brokers (like CommSec) when institutional demand at the asking price isn’t fully absorbing the float. When articles promoting Anthropic’s IPO to Australian investors suggest retail access is a ‘privilege’ rather than a function of distributing massive, overpriced floats, the authors ignore the historical dynamics of IPOs.
Roger Montgomery is the Founder and Chairman of Montgomery Investment Management. Roger has over three decades of experience in funds management and related activities, including equities analysis, equity and derivatives strategy, trading and stockbroking. Prior to establishing Montgomery, Roger held positions at Ord Minnett Jardine Fleming, BT (Australia) Limited and Merrill Lynch.
He is also author of best-selling investment guide-book for the stock market, Value.able – how to value the best stocks and buy them for less than they are worth.
Roger appears regularly on television and radio, and in the press, including ABC radio and TV, The Australian and Ausbiz. View upcoming media appearances.
This post was contributed by a representative of Montgomery Investment Management Pty Limited (AFSL No. 354564). The principal purpose of this post is to provide factual information and not provide financial product advice. Additionally, the information provided is not intended to provide any recommendation or opinion about any financial product. Any commentary and statements of opinion however may contain general advice only that is prepared without taking into account your personal objectives, financial circumstances or needs. Because of this, before acting on any of the information provided, you should always consider its appropriateness in light of your personal objectives, financial circumstances and needs and should consider seeking independent advice from a financial advisor if necessary before making any decisions. This post specifically excludes personal advice.
