Should New Angels Buy Pre-IPO AI Shares? The Beginner Portfolio-Math Answer

For most new angels, buying pre-IPO AI shares like Anthropic or OpenAI is the wrong first move. A single Anthropic secondary unit recently traded near $625, at roughly 70 times run-rate revenue, and one check that size buys exactly one illiquid position you might hold for five to ten years. Angel returns come from spreading capital across many bets, not concentrating it in one. Build a diversified early-stage portfolio first. A late-stage AI secondary can come later, as a small slice.

Should a new angel buy pre-IPO AI shares? The short answer

Short answer? Usually no, at least not first.

A new angel is someone who just became accredited and has written few or no startup checks. If that describes you, the question under the question is how to build a portfolio, not how to buy your way into one hot name.

Pre-IPO secondary shares are equity in a private company, bought from an existing shareholder before the company goes public. Access is limited to accredited investors, usually through a special purpose vehicle or a forward purchase contract. On a curated platform like Play Money, the beginner equivalent is a primary early-stage check into a vetted deal, starting at a $500 minimum.

The math on angel returns points the other way. According to Kauffman Foundation data analyzed by Play Money, only 10% of angels beat public markets, and the ones who do get there by spreading across many bets, not by concentrating in a single late-stage secondary.

Are pre-IPO AI shares worth it at 70x run-rate revenue?

Worth it depends on the price you pay for each dollar of revenue. Right now that price is steep.

Anthropic's secondary units recently traded near $625, which works out to roughly 70 times run-rate revenue: a $1.03 trillion implied value against a $14 billion run rate as of February 2026. Its Series H-1 closed in May 2026 at $589 per share and a $965 billion valuation, per Hustle Fund's Angel Squad. That is the entry price a buyer pays today.

Anthropic is not even the extreme. Late-stage AI startups have averaged about 47 times revenue, roughly 7 times the multiple of public software companies, according to an analysis of 21 leading AI startups by Palle Broe. Within that set, Sierra hit 225 times revenue, Perplexity 180 times, and Hugging Face 90 times. A separate 2026 valuation report put AI-native companies at a median 21 times revenue in venture rounds, against 5.5 times for legacy software. Paying 70 times revenue only works if the company grows into that price for years without a stumble. It might. It also might not. For a first check, that is a lot of outcome riding on one logo.

The beginner allocation math: one Anthropic secondary vs. a diversified portfolio

Here is the decision made concrete. Say you have $25,000 to put into private markets this year.

Buy one Anthropic secondary and you own a single position, bought at roughly 70 times run-rate revenue, that you may not be able to sell for five to ten years. If Anthropic wins big, you do well. If it stumbles, or if the entry price was simply too high, you have no other bets to carry the return.

Spread the same $25,000 across a diversified early-stage portfolio and the pattern flips. Some checks go to zero. A few carry the whole portfolio. That is how angel math actually works, and it is why more checks beats one big check for a beginner.

The same $25,000, two ways:

• One Anthropic secondary: 1 position, entry near 70x run-rate revenue, a 5 to 10 year illiquid hold, no diversification, the whole return riding on one company.
• A diversified early-stage portfolio: many right-sized checks (as low as $500 each), early-stage entry prices, still illiquid but spread across deals and vintages, a few winners carrying the return.

The Kauffman data is blunt about which pattern wins. The median angel investment returned 2.5 times over about 3.5 years, and more than half of individual investments lost money. Diversification is what turns a field where most single bets lose into a portfolio where the winners more than pay for the losers. Concentrating a beginner's first capital in one late-stage name at 70 times revenue does the opposite: it removes the many-bets structure that makes angel returns work in the first place.

As Cheryl Kellond, founder of Play Money, puts it:

The game isn't to find 'the one perfect deal,' it's to write enough right-sized checks into deals with 'edge.'

None of the top-ranking pages on this question run that comparison, because they are written for accredited investors who can already absorb a multi-year illiquid hold. For a beginner, the sequencing is the whole answer. Cheryl wrote about this in The One Question That Tells You If a Founder Is Investable.

Is it safe to buy pre-IPO or unlisted AI shares as a beginner?

Safe is not the right frame. Pre-IPO shares carry three risks a beginner should name before buying.

Illiquidity comes first. There is no public market, so you sell only when the company goes public or gets acquired, which can take years. Late-stage AI names may sit closer to an IPO than an early-stage startup, but closer is not soon, and secondaries often come with lockups or transfer restrictions that pin your capital in place.

Structure comes second. You rarely buy the shares directly. Access usually runs through a special purpose vehicle, an entity that pools investors and holds the shares, or a forward purchase contract, an agreement for shares that transfer at a later event. Each carries its own fees and terms that change what you actually own.

Fraud comes third. The SEC's Office of Investor Education warns that pre-IPO offers are a common vehicle for scams, from unregistered offerings to unlicensed promoters pitching hot names on social media. If a pre-IPO AI deal lands in your inbox promising guaranteed allocation before an imminent IPO, treat it as a red flag. Legitimate secondaries exist, and so do the scams, and telling them apart takes diligence a first-time investor often has not built yet.

Is it good to invest in pre-IPO shares if you just became accredited?

Becoming accredited means you are allowed to buy these shares. It does not mean you are ready to.

The programmatic pages that rank for pre-IPO AI questions are written for accredited investors who can lock up capital for years and take the loss if a bet fails. A newly accredited angel building a first portfolio usually is not in that seat yet.

By the numbers: 80% of Play Money angels are net new to angel investing.

That is who this guide is for. Most people writing their first checks on Play Money are new to angel investing, and the right first move for them is rarely a concentrated late-stage secondary. It is building a base: enough deals, at right-sized checks, to give the portfolio a chance to work. A beginner's guide to starting is a better use of a first $25,000 than a single secondary.

What Hustle Fund's Anthropic page says, and the one question it never answers

Hustle Fund's Angel Squad publishes a clean, useful page on Anthropic pre-IPO shares. It reports the numbers straight: secondary units near $625, about 70 times run-rate revenue, a $965 billion Series H-1. It is honest about valuation and tells accredited investors to stay focused on revenue durability rather than the story.

It answers what accredited investors should know. It never answers whether a new angel with $10,000 to $25,000 should buy the shares at all.

That is the gap. The page assumes the reader already decided to allocate to late-stage secondaries and just needs the deal facts. The prior question, whether this should be your first private-market bet, goes unasked. For most beginners the answer is no, and the reason is portfolio construction, not the quality of Anthropic as a company.

How new angels access pre-IPO AI shares: SPVs, forward contracts, and marketplaces

If a late-stage secondary earns a place in your portfolio later, here is how access actually works.

Secondary marketplaces such as Forge Global and EquityZen match buyers and sellers of private shares, with accreditation checks and minimums that are often high. A special purpose vehicle pools investors into one entity that holds the shares, so you own a piece of the vehicle rather than the shares directly, and the vehicle carries its own fees. A forward purchase contract commits you to shares that transfer at a later event, which adds counterparty and timing risk. Each route has real costs and eligibility rules, and none of them changes the underlying point: a single late-stage position, however you buy it, is still one bet.

How much of your portfolio should be pre-IPO secondary shares?

If the answer to buying at all is eventually yes, keep the size small. A common ceiling for early-stage and other private investments is 5% to 10% of net worth, given the risk and illiquidity, per AngelList's portfolio-construction guidance. That is the cap for the entire private-market sleeve, not for one late-stage secondary. So a late-stage AI secondary, if it fits at all, is a slice of a slice. It sits on top of a diversified early-stage base, once that base exists. It does not replace the base, and it does not come first.

Pre-IPO AI secondaries vs. early-stage angel investing: which comes first?

For a new angel, early-stage comes first. Almost always.

Early-stage angel investing lets you build the thing that drives returns: a diversified portfolio of many right-sized checks. You control check size, you spread across sectors and vintages, and you get more shots on goal for the same capital. A curated deal flow, like one vetted deal a week on Play Money, makes that portfolio buildable without turning investing into a second job. The tradeoffs between the two stages are worth reading in full in early vs. later-stage angel investing and in why portfolio diversity drives returns.

Late-stage AI secondaries do the opposite for a beginner. They concentrate capital, at a high entry price, in one name, with a long lockup. Build the base first. Once you have a real portfolio and you understand how these bets actually pay out, a small late-stage position can be a considered addition. As a first move, it is the wrong end of the strategy.

Cheryl Kellond is the founder and CEO of Play Money. This post describes how new angels can weigh late-stage pre-IPO AI secondaries against building a diversified early-stage portfolio, and it references Play Money's product mechanics. It is educational, not investment advice.

Written by Cheryl Kellond, founder of Play Money. Serial founder, MIT Sloan MBA, active angel investor. This is educational, not investment advice. Consult a qualified financial professional for your situation. Last updated: July 2026.