ResearchHub is an open-science platform that pays researchers in its own token, ResearchCoin (RSC), for work the academic system normally extracts for free – above all peer review. It is the most-cited working instance of the paid-review thesis described in decentralized publishing and peer review, and it is unusual among DeSci organizations in that it is not an IP-NFT funding vehicle at all: it sells no research assets and owns no patents. Its product is the reviewing and publishing layer itself.
It is also the DeSci project with the most direct line to a public company. ResearchHub was founded by Brian Armstrong, the chairman and CEO of Coinbase, and RSC now trades on Coinbase – an arrangement Coinbase discloses in its own proxy statement as a related-party transaction, quoted in full below.
This page describes what the network verifiably is, which is narrower than how it is often described. ResearchHub is frequently called a DAO. Its own documentation says it functions “like” one. On inspection the RSC contract has no voting machinery, votes are non-binding signals, and execution runs through small multisigs – a gap this page documents in detail.
The idea was published on 25 February 2019 in Armstrong's essay “Ideas on how to improve scientific research”, which argued that science's core problems were incentive problems: reviewers work unpaid and uncredited, publishers capture the value, and replication earns nothing. RSC was proposed as the missing reward unit. Armstrong later described ResearchHub as “a project I co-founded with Patrick Joyce,” and noted that “a separate foundation, the ResearchHub Foundation, created ResearchCoin.”
The structure has three parts that are easy to conflate:
Fundraising is documented in SEC filings rather than press releases, and the filings are the more precise source. The 2023 round offered $5,266,546 and sold $4,462,576 to four investors, led by OSS Capital; a second round sold $2,000,000 to two investors from February 2025. Coverage describing a clean “$5M raise” is quoting the offering size, not the amount sold.
RSC is an ordinary fixed-supply ERC-20. The Ethereum contract (0xD101dCC414F310268c37eEb4cD376CcFA507F571) reports a totalSupply of exactly 1,000,000,000 RSC, fully pre-mined at the 1 August 2020 token generation event – there is no mint function. The Base contract is a bridged representation, “backed 1:1 by the main token on Ethereum”; the two balances are not additive, so the true supply is 1B, not the ~1.08B a naive sum suggests (the Base contract's totalSupply read 84m on 29 September 2026).
The contract carries no delegate, getVotes, or checkpointing functions – none of the ERC20Votes machinery that on-chain token governance requires. This is the technical reason RSC cannot cast a binding on-chain vote, and it is a deliberate design, not an oversight: RSC is a reward and reputation unit.
Distribution is documented as “60% distributed to the community and users of ResearchHub, with a maximum emission of 5% per year,” “20% retained by ResearchHub Technologies, Inc.,” “10% retained by ResearchHub's founders,” and “10% reserved for future ResearchHub employees.” The Foundation “may receive 1% per year (10M RSC)” of the community emission.
The 60% community share is a schedule, not a present-tense fact, and market-data sites do not measure how far it has run. CoinGecko has reported about 221m RSC circulating, then 999.8m on 29 September 2026, then about 243m on 18 October 2026. Emission is capped at 5% per year, so full community distribution is a decade-scale process by construction. RIP-24, passed in January 2026, set that year's release at 4%, 40m RSC: 10m for peer review, 20m for research funding and 10m for partnerships with universities and research organizations.
RSC is earned through peer review, bounties, and (historically) upvotes, and is spent on funding proposals, bounties, and journal fees. ResearchHub takes “a 9% fee on any amounts funded to proposals”.
RIP-23, passed in August 2025 with 2.04m RSC for and none against, commits the Foundation to burn every RSC it collects in platform fees, which the proposal put at 2% on funding and bounties, by sending it each week from its hot wallet to the null address. Read on 29 September 2026, the null address holds no RSC on either chain, and the Base contract's 0x…dEaD address holds about 200,000.
The paid-review program is ResearchHub's most concrete contribution and the part most worth studying. The published guidelines are specific: “$150 USD equivalent in ResearchCoin (RSC) per approved high-quality review”, with “a maximum of two paid reviews per paper,” reviewers limited to “up to 2 reviews per week,” reviews of “up to 3,000 words,” and payment “approximately 10 days after acceptance.”
The sybil defence is identity, not stake: to be paid, reviewers must “verify their identity” and “authenticate ORCID.” Notably, no PhD or publication record is required – reviews are editorially vetted before payment rather than gated by credential beforehand. This is the inverse of the traditional model's trust assumption, and the honest open question is whether editorial vetting scales as well as reputational gatekeeping.
The ResearchHub Journal (ISSN 3070-3395) applies the model to a full journal. At launch it posted every submission as a preprint, promised peer review in about 14 days and a decision in about 21, and charged a $1,000 article processing charge. The journal page, read on 18 October 2026, describes a narrower journal: a multidisciplinary registered-reports journal open only to studies funded on ResearchHub, with open, unblinded and paid peer review, optional monthly progress updates from authors, a CC-BY 4.0 licence and no processing charge.
Its output has been small. A query of ResearchHub's own journal feed in July 2026 returned 12 papers in total since launch, spanning July 2025 to April 2026, with nothing published in the roughly three months to July 2026: well under one paper a month. The journal is also not yet indexed in DOAJ or OpenAlex; ResearchHub's documentation promises indexing only in the future tense.
ResearchHub is routinely listed among DAOs, and its documentation says it functions “like a decentralized autonomous organization,” with “token holders of ResearchCoin voting on proposals.” That second clause overstates what exists, and the gap is instructive for anyone studying on-chain versus off-chain governance.
What is verifiable from the researchhub.eth Snapshot space and the token contract:
erc20-balance-of on the Ethereum and Base contracts – plain token-weighted voting in signalling form. Across the space's entire history there have been 29 proposals.quorum is unset.onlyMembers: true with three admins and one member – four addresses may propose. It is not open proposal-making.Turnout reflects this. The most recent proposal, RIP-24 (19 January 2026), drew five voters casting about 53,000 RSC (36,000 for and 17,000 against) – against a supply of a billion. ResearchHub's own team acknowledged the underlying design early: on Hacker News in 2021 a team member wrote that “RSC is 100% pre-mined” and “for every coin held, an individual wallet gets one vote. This means that the biggest hodlers of RSC will have the most influence.”
The accurate description is therefore: off-chain, non-binding, balance-weighted signalling with gated proposal rights, no quorum, and multisig execution. That is a legitimate way to run a company-led open-science platform – but it is a company with a suggestion box, not a DAO in the token-voting sense, and the wiki records it as such.
The Coinbase relationship is disclosed, and the disclosure is the story. Coinbase's DEF 14A proxy statement filed 24 April 2026 lists RSC under related-party transactions, stating verbatim:
“In the first quarter of 2025, ResearchCoin was approved for listing on our retail trading platform by the digital asset support group of Coinbase, Inc. in accordance with our digital asset support policies and procedures. Mr. Armstrong, our Chairman & Chief Executive Officer, is a co-founder and Chief Executive Officer of ResearchHub Technologies, Inc., the creator of ResearchCoin, and holds a significant percentage of ResearchCoin's market capitalization. We recognized revenue of approximately $201,273 from ResearchCoin in the year ended December 31, 2025.”
Two things follow. Coinbase itself concedes that its CEO “holds a significant percentage” of the market cap of a token his exchange lists, and it quantifies its own revenue from it. Coinbase's stated procedural answer is that “Brian is not a reviewing or approving member of the Coinbase Digital Asset Support Group”. Readers should note the timeline runs from Q1 2025 approval to the 31 July 2025 public listing.
Reward farming is self-admitted. The clearest evidence is ResearchHub's own proposal RIP-22 (June 2025), which removed the 0.01 RSC upvote reward, stating that “while rewarding upvotes was originally intended to encourage participation and reward high quality content, in practice it has created an easy opportunity for spam users and bots to extract value from the ResearchHub community.” This is the predictable failure of paying for a cheap, unverifiable action, and it is a compact case study for sybil resistance: the fix was to stop paying for the action that could not be verified, and keep paying for the one that could be reviewed.
The market has been unkind. RSC traded at $0.07 on 18 October 2026, about 95% below its $1.51 all-time high of January 2025; at that price the full billion-token supply is worth about $70m. Because reviewer pay is denominated in USD and settled in RSC, the price collapse does not cut a reviewer's headline $150 – but it does mean the Foundation spends far more RSC per review than it once did, and it weakens the token's story as an appreciating claim on future science.
Two frequent misattributions are worth correcting. The Scholarly Kitchen's well-known critique of paid review is a generic argument that does not mention ResearchHub and predates the journal by years; and critic David Gerard's write-up describes ResearchHub as “founded by crypto exchange Coinbase,” which is incorrect – it was founded by Coinbase's CEO personally, a distinction the SEC filing makes precisely. Substantive independent coverage of ResearchHub remains thin; the filings are better sources than the journalism.
ResearchHub and Caper are solving different problems – ResearchHub pays for scientific labour, which a caper does not do – but they collide on one question: what should a token actually measure?
ResearchHub answers it by splitting the two roles and letting one wither. RSC is a reward unit, so it does its job well: a review is verifiable work, and paying for it in a transferable token is coherent. But because RSC is also the governance instrument, and because it is transferable and 40% of it is retained by the company, its founders and future employees, governance inherits a token that was never designed to carry it – hence no quorum, four permitted proposers, and five voters on the most recent proposal. Influence there is priced; anyone can buy RSC without ever reviewing a paper.
A caper keeps the two units apart. A member's say grows with the caper's tokens they hold and with the votes they have cast, and the votes sit in a soulbound token that cannot be sold, sent or delegated: a member earns it by casting a ranked ballot, and buying or selling earns none. The same measure decides a member's share of the treasury when they exit, so voice and exit move together.
Holdings still count, and a large position matters. They are one of two inputs, though, and the other is a record of the member's own ballots, which nobody can buy from them, so a big bag alone can't capture control. ResearchHub shows the value of paying for work that can be checked; a caper applies the same instinct to governance.
ResearchHub's notion of contribution is far richer than a caper's. It knows a good review from a bad one, because a human read it. A caper knows only whether you voted. That is a much poorer signal – deliberately so, because it is the only one a contract can verify without a judge.
See also: Decentralized publishing & peer review, DeSci DAOs, Molecule, Bio Protocol, and Token-weighted voting.