The 25-Minute Stock Market: Robinhood’s First Two Months Onchain

Introduction

Robinhood Chain launched on 2 July 2026 with 194 tokenized equities, ETFs and commodities issued on the brokerage’s own Layer 2. It was the largest retail brokerage entry into tokenized equities to date, and it arrived last into a category already served by xStocks, Ondo Global Markets, bStocks, Backpack Securities and PreStocks.

Sixty-one days later, the chain had processed 23,336,757 swaps across 215,775 wallets, cleared $3.66 billion in trading volume, and accumulated more than one million holder accounts. Within weeks of launch it was among the most heavily traded venues in the category, on a book far smaller than those of the issuers it trades alongside.

This report examines all onchain trades of the 194 tokens from launch through 1 September 2026, together with a 4 September holder snapshot, to establish how long positions were held, what the tokens were traded against, how much of the volume represented capital changing hands, and how ownership is distributed.

A small book heavily traded

On Robinhood, market capitalization expanded from $1.24 million on the first day of trading to $66.2 million by 1 September, a 53-fold increase across sixty-one days, and reached $107.9 million by 4 September. The book remains small against other issuers, but on trading volume Robinhood is outperforming.

Trading volume rose from negligible activity in early July to a single-day record of $703.7 million on 1 September — 4.8 times the previous high. That record came at the close of a nine-day run that advanced without a single down day. Cumulative volume over the two months reached $3.66 billion across 23.3 million trades. Holder accounts moved in step, growing from zero to 1,012,005 across 193 live products, the curve close to flat for six weeks before turning near-vertical through the final fortnight.

The three are not independent measures, and they share a shape. Each turns sharply upward in the same closing stretch: 62% of the market capitalization, 65% of the trading volume and 35% of the holder accounts recorded across the entire period arrived in its final nine days. Most of what Robinhood Chain has produced to date happened at the very end of it.

The median position lasted 24.6 minutes

All three of those measures record activity rather than ownership. Whether the activity amounts to investment depends on how long positions are kept, which the trade data answers directly. Across all matched buy-and-sell activity, the median position on Robinhood Chain was held for 24.6 minutes when weighted by dollar value. 34.5% of matched dollars were sold back within one minute, 56.3% within one hour, 83.5% within a day and 96.7% within a week.

The distribution is short at both ends. The fastest quartile exited within a single second, indicating same-block automated activity rather than discretionary trading. At the opposite end, the 90th percentile exited within two and a half days and the 99th percentile within 20.4 days. Only 3.3% of dollars were held longer than one week, 1.6% longer than two weeks, and 0.3% longer than one month.

Splitting the 19.3 million matched round trips by asset type, counted per position rather than by dollar value, shows the same behaviour everywhere and one instructive exception. Tokenized stocks carry a median round trip of 18.3 minutes, with a quarter of positions closed inside 24 seconds. ETFs run longer at 45.5 minutes and commodities shorter at 15.6 minutes.

Only US Treasuries behave like a held asset, at a median of 5.2 hours with the top quartile stretching beyond two days, roughly seventeen times the stock median. They also account for 12,123 round trips, 0.06% of the total. Weighting by dollar value widens rather than narrows that gap: the Treasury median stretches to 18.7 hours against 25.0 minutes for stocks, meaning the largest Treasury positions are the ones held longest, while in every other category the largest positions close faster than the smallest.

Those durations cover positions that closed. Of all dollars ever used to buy a token on the chain, 85% had been sold back by the end of the period and fall within the distribution above; the remaining 15% was still held and does not.

Positions also tended not to close against the asset they opened against. 90.7% of dollars — $3.32 billion of the $3.66 billion traded — belong to wallet-token positions that touched more than one quote asset, entering against a stablecoin and exiting against something else. The pattern is consistent with inventory moving between venues rather than directional positioning. The 9.3% that remained within a single pairing is examined below.

Memecoin pairings rebounded after a single-week trough

Tokenized stocks on Robinhood Chain trade against four categories of counter-asset: stablecoins, principally Robinhood’s own USDG; memecoins; ETH; and other tokenized stocks. The memecoin-paired share of volume peaked near 46% in late July, fell to 16% in the week of 17 August, and then rebounded to 31% and subsequently 38%.

The memecoins paired against each equity are thematically matched to the ticker they trade beside: AI against NVDA ($105 million), GME against GME ($49 million), BONER against HIMS ($27 million), MOO against MU ($26 million), SPACEHOOD against SPCX ($22 million) and SAYLORMOON against MSTR ($18 million).

Pool infrastructure reflects the same pattern. 122,000 trading pools have been created on the chain and 46,258 have recorded a single trade; the median pool was active for less than one day, and the top 100 pools carry 70% of all volume.

Round-trip trading accounted for 91% of volume

Of the $3.66 billion in gross volume, $3.33 billion — 91% — was round-trip: dollars used to buy a token and subsequently sell it back, leaving no lasting position. The remaining $330 million represents net-directional flow, where a wallet ended the period holding more or less than it began with. Weekly net flow ran at approximately ±$5 million against gross weeks in the hundreds of millions.

A central test of whether a distinct buy-and-hold market exists beneath the speculative activity is whether stablecoin-paired trading — the pairing most closely resembling conventional equity purchase — behaves differently from memecoin-paired trading. It does not. Restricting the sample to wallets that remained within a single pairing for a given stock, the cleanest available view, stablecoin-paired trading round-trips 85% of its dollars, stock-against-stock 83%, ETH-paired 69% and memecoin-paired 64%. Every pairing cycles more than 60% of its dollars, and the stablecoin lane cycles hardest of the four. Stablecoin and stock-against-stock pairings are also the most bot-concentrated segments on the chain.

Those four figures describe the 9.3% of flow that stayed within a single pairing. The 90.7% that crossed pairings round-trips 92%, and the two groups together produce the 91% chain-wide figure.

Activity is heavily concentrated in a small number of wallets. 86% of all trades originate from the 28,000 wallets with 100 or more lifetime trades, and more than half from approximately 2,900 wallets. Among the most active, 157 coordinated groups trade identical sets of pools, the largest a single cluster of 99 wallets under one operator. These trade in bursts at varied sizes rather than on a fixed schedule, a profile consistent with automated market-making. The median wallet made ten trades across a 1.3-day active span, and 6.3% bought once and held.

One wallet holds most of the median token

The 4 September holder snapshot shows ownership more concentrated than trading. For the median tokenized stock, the single largest wallet holds 56% of supply and the top ten hold 90%. For 108 of 193 tokens, one wallet holds more than half the supply.

This is not solely an artefact of liquidity pool and custodial addresses. Removing each token’s single largest holder and renormalizing, the next nine wallets still hold 76% of the median token, and 44% of tokens remain more than 80% held by their top wallets.

The holder base is narrow outside the largest names. The median token has 88 holders, and only 26 of 193 tokens have more than 10,000. Concentration falls as the holder base widens but remains high throughout: for tokens with more than 10,000 holders the largest wallet holds a median 35% and the top ten hold 85%; for tokens under 1,000 holders those figures are 66% and 92%. The flagship products are no exception — NVIDIA, with 103,000 holder accounts, has 74% of supply in its top ten wallets, and SPY’s single largest wallet holds 49%.

Holdings per wallet are correspondingly small. Combining holder counts with market value, the typical holder outside a token’s top ten holds approximately $21 of tokenized stock. NVDA works out at roughly $36 per holder, SPY at $86, AAPL at $17 and AMD at $4.6. Note that these are per-token averages rather than observed balances: the snapshot returns each token’s ten largest holders, so what the remaining holders own is inferred from the supply left over.

Conclusion

Nothing in the two months of trade data points to a durable buy-and-hold market forming beneath the speculative activity. What the data describes instead is a continuously operating trading venue in which tokenized equity tickers function primarily as quote assets and intermediate steps rather than as held positions. Volume, market capitalization and holder counts have all grown quickly against a book that remains one of the category’s smaller ones, and all three are measuring the same behaviour: positions opened and closed within the hour, cycled by a concentrated set of automated participants, against tokens whose supply sits in ten wallets apiece and whose ordinary holders own a few dollars each.

That pattern is no longer confined to Robinhood Chain. Pairing equity tickers against thematically matched memecoins has begun appearing on other chains, which suggests the category is reading the launch as a template rather than as a quirk of one venue. Sixty-one days is a short record on which to settle what any of it means: the measures that would revise the reading — net-directional flow rising with gross volume, positions closing against the assets they opened against, ownership dispersing below the top ten wallets — are all ones that need quarters rather than weeks to show a direction. The findings describe Robinhood Chain through 1 September; where the model settles, on this chain or the ones now copying it, is a question for a longer window.