Crypto people say it like a confession: retail only shows up when they think they can get rich. It usually arrives as a knock on retail, proof the space is unserious, casino-brained, a permanent floor of degens.
I think that reading is backwards. Retail is behaving rationally. If you can make money somewhere you go there, and if you can’t you leave. That isn’t a character flaw, it’s how every market works. The question worth asking isn’t why retail chases money. It’s why crypto keeps losing them, and what 2025 showed when it ran the same play twice, once with tokens and once with stocks.
Start with the boring part
Retail participation in crypto tracks price, not utility. Global retail transaction volume fell 11% year over year to $979B in Q1 2026 (TRM Labs), in lockstep with a softer market. Ownership keeps drifting up, somewhere around one in four US adults, roughly 67 million people, but the activity inside those accounts follows the chart. People hold through the quiet and trade when the candles are green. New buyers arrive after a rally, not before one.
This is ordinary feedback trading, and it isn’t unique to crypto. Crypto just amplifies it. A stock has earnings; a bond has a coupon. Most tokens have neither, so price action, scarcity, and narrative carry the entire weight of “why own this.” Strip out cash flow and the loudest signal left is the price itself.
The obvious objection is that retail leaves for other reasons: regulation, complexity, bad UX. But 2024 and 2025 knocked those barriers down hard. Spot ETFs, clearer rules, custody that finally works. Ownership still only crept. Lower barriers were never the missing piece. Winning was.
2025 ran the experiment
Last year handed us two VC-exit launch markets to compare. Jay Ritter counts about 90 US operating-company IPOs in 2025; counted more broadly, EY tallied 1,293 IPOs globally raising $171.8B. Alongside that, a wave of VC-backed token launches hit exchanges.
Both are versions of the same machine. Private capital builds an asset for years, then sells a slice to the public at the moment of peak attention.
But they are not the same population, and that matters before any comparison. To IPO, a company clears underwriters, SEC review, and audited financials; the roughly 90 that made it are survivors of a filter. A token generation event clears almost none of that. So part of any gap between the two is curation, not structure: the honest comparison to 118 VC-backed tokens isn’t 90 IPOs, it’s the full field of VC-funded companies, most of which fail quietly before they ever file. Hold that thought. The structural argument is what’s left standing after you account for it.
What they share: the pop is the point
Both are built around a launch pop, and around who is allowed to catch it.
IPO underpricing is deliberate. Underwriters price the deal below where the stock will actually open, so the buyers allocated at the offer price get an instant gift and the deal “works.” Those allocations go to institutions on the roadshow. Retail is mostly not in that room; it buys later, on the open market, into the pop. From 1980 through 2025 the average IPO popped about 19% on day one. Circle priced at $31 in June 2025 and closed its first session near $83. The fund allocated at $31 caught all of that. Retail buying at $70 or $90 on day one caught a fraction, or none.
Tokens run the same play with a different lever. Launch a thin slice of supply, let price discovery happen on that thin float, and the headline valuation prints enormous. Insiders and early funds are in cheap; retail buys the pop; the headline number becomes the marketing for the next deal.
Where they split: what happened to the buyer who held
Here the analogy breaks, and the break is the point.
The 2025 IPO class mostly rose after the pop. Circle ran from its $31 offer past $200 within weeks, though most of that move came after the US Senate passed stablecoin legislation, an outside catalyst rather than anything structural. Figma debuted up more than 250%. CoreWeave is up triple digits since listing. As a group, 2025 IPOs that raised more than $50M returned well past 40%. A retail buyer who bought these in the open market after the pop, not at the offer price they couldn’t get, still mostly came out ahead in this cohort.
A couple of caveats keep that honest. This was an unusually hot IPO tape; the same market in 2022 handed retail losses across the board. And the window is short — Ritter’s long-run data shows the average IPO underperforms the market over three years, so “green after a few months” is not “green for good.”
VC tokens gave the opposite result, over a longer window. Of 118 VC-backed tokens launched in 2025, around 85% now trade below their launch price, with a median drawdown near 71% on a fully diluted basis. Roughly 60% fell below even their last private fundraising valuation within six months (Messari). Same launch mechanics, opposite outcome — and that token verdict is measured months-to-a-year deep, not at a few-week snapshot.
Why: float and the claim
Two structural reasons, and they survive the caveats above.
The first is float and unlocks. The median 2025 token launched with a market-cap-to-FDV ratio around 12%, meaning roughly 88% of supply sat locked behind multi-year vesting. Across 2025, about $97B of tokens unlocked into the market, with single weeks carrying more than $700M of scheduled supply. That is a wall of sell-side flow with no matching demand attached to it. An IPO has a 180-day lockup and then a bounded, largely one-time jump in float. A VC token has a release schedule that runs for years, front-loaded and fixed and public, so the market can see the seller coming. It does end once vesting completes. But for the first few years the scheduled supply routinely dwarfs organic buying, and everyone can read the calendar.
Equity isn’t dilution-free either. Companies issue stock continuously through stock-based comp. The difference is that token dilution is front-loaded and runs on a fixed timetable traders can position against, while equity dilution is slower and discretionary. Same force, sharper edge.
The second is the claim. An IPO share is a claim on a real company’s earnings, reported quarterly, audited, with the option to retain and reinvest. Most tokens are deliberately built not to be that, to stay clear of securities law. I’ve argued before that this is a solvable problem and that default tokenomics still deserves the bearish case. Remove the cash-flow claim and the only thing holding a price up is the next buyer. When the unlock calendar guarantees a larger seller every week, the next buyer is structurally outgunned.
Crypto rents retail. It never owns them.
Put it together. Retail comes for the money — confirmed, and reasonable. In 2025 one launch market handed its buyers a cohort that mostly rose, and the other handed them a cohort that mostly sank.
The token didn’t sink because the team was crooked. The structure produces that outcome on its own, whatever anyone intended: when most of the supply is locked and scheduled to release on a public timetable, the small float that early buyers can actually trade is priced against a known, larger future seller. You don’t need bad faith for that to end badly. You need the math.
And this isn’t unique to VC tokens. Fair-launch tokens and memecoins go to zero at least as reliably; speculative crypto assets broadly don’t compound. What’s specific about the VC token is the costume. It arrives wearing a fundraising valuation and an investor cap table, dressed as an investment, while behaving like the rest of the casino. The disappointment is sharper because the framing promised more.
That is why crypto keeps losing retail. Not barriers — those came down. Not attention — launches get plenty. The median token leaves the median holder underwater, and retail, being rational, eventually stops showing up.
The next real retail wave won’t come from a smoother onramp. It comes from tokens whose post-launch structure works the way the 2025 IPOs’ did for the people who held them: float that isn’t a countdown to dilution, and a design where holding the asset is a claim on something real rather than a seat ahead of the next unlock. Revenue-funded buybacks are one version of that. Conviction-aligned vesting is another.
Until default token structure changes, every cycle repeats the same move: it rents retail on the way up and never gets to own them. The 2025 IPO market kept retail’s attention for one boring reason: the asset mostly went up. Not because retail got the insider’s allocation; they didn’t. Because the thing they were allowed to buy wasn’t built around a seller bigger than them. Most tokens are. That’s the whole formula, and most token designs still get it backwards.
