Six Years and $2.2M Later, Brian Armstrong's Science Platform Is Paying Off

Introduction

ResearchHub has deployed more than $2.2 million into scientific research, pays $150 for every accepted peer review, and gets those reviews back in under ten days on average. Over 100,000 researchers and readers use it. Brian Armstrong co-founded it in 2020, and the work has been funded since 2023 by a $5M round led by Open Source Software Capital. Six years in, it's the rare decentralized science project running a real product loop rather than a thesis. DeSci as a category has been through the wringer since the January 2025 top, and ResearchHub kept shipping straight through it. The platform is doing more today than it was when the sector was worth eight times what it is now.

What the product actually is

Traditional grantmaking wants a long application, takes months or years to review, and routes a large share of the award to the university as overhead. ResearchHub compressed that into four steps. A researcher publicly preregisters a two-to-four page study design covering background, hypothesis, methods, analysis plan and budget; reviewers critique it in the open and get paid for it; funders commit capital against the proposal; results get linked back to the preregistration whether they're positive or negative. Indirects are held between 0% and 10%, and money reaches the lab as a tax-deductible gift routed through Endaoment, the 501(c)(3) that converts RSC and handles custody. Labs at Stanford, Harvard, Cornell, Purdue, UC San Diego, Indiana and Iowa State have all taken funding this way.

The preregistration piece is doing more work than it looks. Less than half of study outcomes ever get reported and less than half of those replicate, and locking a design in public before data collection is the cheapest available fix; in other words, the funding mechanism and the reproducibility pitch are the same mechanism. Funded proposals can then publish as registered reports in the ResearchHub Journal, which was recently made free for any proposal that gets funded, and turns review around in as little as 14 days.

Those steps are increasingly run by software rather than staff. An AI Expert Finder sources applicants for open RFPs, AI reviews generate on proposals instantly, and an AI Research Assistant is about to go live, with several of these opening to the public shortly. Stacked together they point at a mostly automated path from proposal to matched reviewers to funded study to published result, running on the sub-ten-day clock the platform already holds against grant cycles that take months.

Paying for peer review is the wedge

Peer review is the load-bearing unpaid labor of the entire scientific publishing industry, and it's the part ResearchHub went after first. Reviewers get $150 in RSC per accepted review, ORCID verification required, priced in dollars at the moment of distribution so the payout doesn't depend on where the token trades that week. Volume is real and recurring: 1,401 reviews completed in Q1 of this year alone, from researchers with no particular interest in crypto.

Nature covered the platform in December 2024 and found São Paulo molecular biology consultant Pedro Paulo Gattai Gomes reviewing roughly 15 papers a month and earning more from it than he had in his academic posts. Getting paid to review, he told Nature, is "justice."

The endowment turns holding $RSC into a research budget

This is the newest piece and the most crypto-native thing on the platform. Deposit RSC into a ResearchHub account and you accrue daily Funding Credits, which can only be spent funding preregistered proposals. Principal stays withdrawable with no lockups, and holding age earns a ve-style multiplier: 1x under 30 days, 1.05x past 30, 1.1x past 180, 1.25x past a year.

The yield is protocol emission, 9.5M RSC in year one on a schedule that halves every 64 years, distributed pro rata. That's dilution rather than revenue, and the design routes it somewhere unusual. A Funding Credit can't be sold, so the emission never reaches the market as a holder exit; it converts only by funding a study, at which point Endaoment sells the RSC down to move dollars to the university, and 2% of every funded amount is burned on the way through. Every points program of the last two cycles ended in an airdrop dumped inside a week. Here the only sell pressure the reward creates arrives at the moment an experiment gets paid for, and a slice of supply disappears each time it happens. 

What the money has actually produced

The clearest result so far came out of Guy Fincham's Breathwork Lab at Brighton & Sussex Medical School. His Airways to Alteration trial, funded by ResearchHub alongside the nonprofit DMT Quest, went from preprint in November 2025 to publication in Frontiers in Psychology on June 10, 2026. Fincham isn't a fringe researcher chasing fringe money; his 2023 breathwork meta-analysis in Scientific Reports has been cited over 350 times.

The open RFPs are stranger and considerably more fun. There's up to $100,000 to test whether low-frequency EMF exposure explains why the 49ers are the NFL's most injured team, and $200,000 from a pseudonymous donor to test Joseph Davidovits' claim that ancient megaliths were cast from geopolymer rather than quarried. Uma Chatterjee's OCD proposal pulled in 29 separate funders. These are hypotheses no NIH study section would go near, which is the entire argument for building a parallel funding rail, and every one of them will publish its results in public whether they confirm the hypothesis or bury it.

Where it's headed

ResearchHub frames its own edge as automation: preregistrations instead of full grant applications, automated recruitment of proposals and reviewers, indirects under 10%. The roadmap points toward more AI tooling across that pipeline, with the end state being a largely automated path from proposal to review to funded study to published result. That direction is already going institutional; AAAI 2026 ran a frontier reasoning model across 30,948 submissions as a supplement to human reviewers, with humans keeping every decision.

The platform’s science budget is a 455.7M RSC pool sitting in a multisig. It’s minted but earmarked for the community, going out as ecosystem incentives: endowment funding credit yield, peer review bounties, and the rest of the reward surface, released at up to 5% of supply (50M RSC) a year.

Scientists who earn RSC through reviews and bounties have started rolling it into endowments of their own, turning review income into a standing budget for their lab and making the same person a reviewer, a funder, and a grantee. The platform also takes fiat money through card, Apple Pay, or through a donor-advised fund via Endaoment with a tax receipt attached. None of these routes require the funder to hold RSC, meaning ResearchHub applies to the broader scientific community rather than just crypto natives.

Wrapping Up

There's plenty still unproven. $2.2M in cumulative research funding is roughly what one standard NIH R01 pays out over its life, funded teams are incentivized to post updates but carry no formal reporting obligation, the endowment yield is emission rather than revenue, and most of the supply has yet to be distributed, with 238M circulating against a community pool that releases at up to 50M RSC a year.

What's hard to argue with is that the loop closes. A scientist posted a proposal, strangers funded it in RSC, a university foundation received the cash, the experiment ran, reviewers got paid, and the paper came out in a Frontiers journal with ResearchHub in the funding line. Six years into decentralized science, very few projects can produce that sentence.


Liked this article and the ResearchHub vision? Get involved here.