Why AI valuations look extreme on paper

Valuing an artificial intelligence company forces a choice between two worlds. In one world, you anchor to current revenue and earnings, which makes almost every AI leader look absurdly expensive. In the other, you project the growth forward several years, which makes the same company look cheap if it keeps compounding. Most of the market has chosen the second world, and that is why the biggest AI names trade at multiples that would have looked irrational a decade ago.

Anthropic is the clearest current example. Alexander Green, chief investment strategist at The Oxford Club, has made the company the centerpiece of his “Secret Backdoor to the REAL #1 Tech IPO of 2026” pitch in The Oxford Communiqué. The numbers he cites are striking. Revenue climbed from about $1 billion in early 2025 to $11.5 billion in a single quarter, and Bloomberg reported an annualized run-rate of $65 billion by the end of July, up from $47 billion in May.

The $2 trillion question

Green floats a $2 trillion IPO valuation for Anthropic. That is a large number to attach to a company still filing a confidential S-1, so the math is worth doing. At a $65 billion run-rate, a $2 trillion valuation implies roughly a 30-times-sales multiple on annualized revenue. Public software companies rarely sustain multiples that high for long, and when they do, it is usually because the growth rate is still accelerating sharply, which pushes the multiple higher, or because the market is pricing in dominance it has not yet proven.

The Wall Street Journal reported an operating profit of $559 million in the most recent quarter, a figure worth taking in context. That is an operating number, not GAAP net income, and it comes from a company in the middle of an IPO process where the incentive to present the best possible optics is real. It is a strong signal, but it is not the same as a clean, audited earnings statement. The same tension shows up in our guide to the best AI stocks.

The premium problem in the backdoor funds

Here is where valuation gets concrete for retail investors. Because Anthropic is still private, most people get exposure through the funds Green names. Fundrise Venture, a closed-end fund trading as VCX, holds Anthropic as its largest position, alongside Databricks and OpenAI. The fund listed on March 19, 2026 at roughly $19 net asset value, and its NAV today is estimated at $25 to $30. The shares trade around $41, which means buyers are paying a 30 to 40 percent premium to the underlying value. Earlier this spring, the shares hit $200, roughly ten times NAV, before lockup expirations pulled them back.

That premium is the single most important number in the trade. When you buy VCX at $41, you are not just buying Anthropic’s growth. You are paying extra for the privilege, and closed-end fund premiums have a habit of compressing when enthusiasm cools. A strong Anthropic IPO could still leave VCX shareholders flat if the premium they paid evaporates. We dig into the fund mechanics in our Anthropic stock explainer and our piece on Anthropic’s valuation.

What a reasonable investor should do

The honest conclusion is that Anthropic’s growth is real and its valuation ambition is high, and those two facts point in opposite directions. Green’s underlying idea, giving retail investors liquid access to Anthropic, Databricks, and OpenAI, is a genuinely interesting one. The execution risk sits entirely in the price. Waiting for the premium on the closed-end funds to compress, or sizing a small starter position rather than a full allocation, are both more measured ways to participate than paying up for a story that the market has already started to price in.

The business model behind the multiple

A valuation is only as durable as the revenue behind it, and Anthropic’s revenue is unusually concentrated in paying business customers rather than free consumers. The company pivoted early toward enterprise tools like Claude Code for software engineering, and that coding business grew from about $500 million to $2.5 billion in five months. More than 300,000 business customers now anchor the top line. Claude has around 350 million monthly actives, fewer than ChatGPT’s more than one billion weekly users, but a paying business customer is worth far more than a free user, and that higher-margin mix is what supports a premium multiple.

Other ways the market prices exposure

The premium problem is not unique to VCX. SK Telecom, trading as SKM, holds an Anthropic stake it bought for about $100 million that may now be worth more than $2 billion. The Destiny Tech100 Fund, trading as DXYZ, and the KraneShares Public-Private AI & Technology ETF, under the ticker AGIX with about a 2 percent Anthropic weighting, offer thinner exposures of their own. Every route into a private AI company carries either a premium, a fee, or dilution, and understanding which one you are paying is the whole job.

Ready to see the research? Click here to access Alexander Green’s report.

NewsletterVetter is an independent publication. We receive compensation from some of the services we review through affiliate links. Nothing on this site is investment advice. Always do your own research.