Fact Machine
Opinion markets let you trade not on what will happen, but on what people will believe. The crowd becomes the oracle, rewarding those who can read collective sentiment.
TL;DR
Betting on what the crowd will believe
Have you ever argued with a friend about who is better: blondes or brunettes? Messi or Ronaldo? Can men and women really be just friends? Well, now you can do more than just shout louder — you can put your money where your mouth is.
A typical prediction market revolves around a simple question: will something happen or not? Did the Fed cut rates? Did a candidate win? Did BTC hit a certain price? You need resolution, oracle, clear rules and often plenty of arguing over the wording. In the end, there has to be a definitive answer: yes, it happened or no, it didn’t.
Then there is:
Fact MachineMoni Score2429WebsiteAt first glance, it may look like a prediction market. But the core idea is different: you are not trading what will happen - you are trading what people will believe.
That is an important distinction.
On Fact Machine, a market doesn’t resolve based on an objective event. It resolves through user voting. First, you express your own opinion. Then, you trade based on what you think the majority will vote for.
So this is not a market for facts. It is a market for social intuition.
Until the very end, the voting result remains hidden, while the trading market stays open. The winners are the people who correctly read the crowd and picked the option the majority eventually chose.
For example, if 75% of users bet on blondes, but the final vote says the majority chose brunettes, then the 25% who backed that outcome get paid.
In December 2025, the team launched a two-week alpha testnet for a carefully selected group of whitelisted users. Even though it was only a testnet, participants were not left empty-handed: the team set aside $50K in rewards.
Then came the second alpha-testnet season, which wrapped up quite recently. Getting in became slightly easier, since users could invite others, but not everyone passed the checks. The team also handed out various prizes: Pokemon Go trading cards, a Mac Mini, a flamethrower (yeah, an actual flamethrower) a piece of silver, merch, a Lehman Brothers 2008 stock certificate and a bunch of other, let’s say, very non-standard prizes. Just scroll through their X from February onward.
In March, physical rewards ended and the format changed. Every week, the team refreshed test balances to $5,000, and for every $1,000 in positive test PnL, users received $10 in real USDC. Payments happened weekly: 50% went directly to the user’s wallet and other half was supposed to be credited to their mainnet balance.
Throughout the testnet, the team kept experimenting:
They tested different market durations: main 24-hour markets, short 30-minute Flash markets appearing throughout the day and one-hour markets.
They experimented with formats: some markets were binary, the usual “you either win or lose” model, while others used a percentage-based scalar model, where the final price is based on user votes.
For example, if 60% of users voted for one option, its final price becomes $0.60. Bought below that and you profit; bought above it and you are in red, buddy.
On July 15, the team moved to the next stage: from alpha testnet to the beta version of the protocol. They removed the free weekly balance refresh and now users need to deposit real USDC on Solana if their balance goes to zero.
Still, the team remains pretty generous: during this phase, every new user receives $10 to try the product. Anything earned on top of that is theirs to keep and withdraw to their wallet.
Small account, serious room
Because with only around 4K followers, its Moni Score is 2429. It is followed by a number of Solana OGs, broader crypto OGs, and people from well-known VCs like Delphi Digital, a16z and Paradigm. That is an unusual follower-to-score ratio.
Another interesting detail: the project is listed on the Colosseum website under the Strategic category. That may suggest the team sees FactMachine as important and could be involved more deeply than usual.
As for the team, it is small and mostly made up of young builders from the US. Their previous experience is mainly in consumer web2 products such as TikTok, Twitch and Meta, but there is also crypto experience: Coinbase and one of the core developers previously worked in the prediction markets niche as an engineer at Noise:
NoiseMoni Score7818WebsiteThere has been no separate public funding announcement, but one of the February job postings for a mobile app developer mentioned that the project had raised an $8.2M round, which they planned to announce closer to mainnet.
jobs.solana.comSolana Network Opportunities Job BoardSearch job openings across the Solana Network Opportunities network.https://jobs.solana.com/companies/fact-machine/jobs/66757122-mobile-app-lead
Where prediction markets hit their ceiling
Prediction markets already proved that event trading has real demand. Polymarket and Kalshi have become part of the everyday information layer: market prices are increasingly treated as a live read on what the crowd expects next.
But prediction markets have a limitation: a good market needs a clearly resolvable outcome.
The real world is messier than that.
Most things people actually argue about online do not have a clean oracle. Who won the debate? Which L2 has more mindshare right now? Was a launch successful? Is this a good take or a bad take? Is Mamdani bullish or cringe? There is no neutral source of truth for questions like that.
Galaxy Research describes opinion markets as markets for subjective questions and narratives, where price reflects collective opinion rather than final objective truth.
This is exactly where Fact Machine finds an interesting opening.
On one hand, it sounds a little unserious. On the other, the internet has already been operating like a giant opinion market for years, just without a proper settlement layer. People already farm attention, reputation, likes, reposts, CT narratives and VC visibility through their takes. Fact Machine simply makes that layer more explicit: if you think you understand the crowd better than others, put money on it.
What happens when the crowd is the oracle?
In a classic prediction market, you can move price with money, but you cannot change the fact itself. The match will end. The Fed decision will come out. The candidate will either win or lose.
In an opinion market, things are more delicate. The participant is not only a trader. They are also part of the oracle.
If resolution depends on voting, the main attack vector is not buying the outcome through the orderbook. It is influencing the vote split itself. That could mean strategic voting, multi-accounting, group coordination or simply voting not for your real opinion, but for the result where you hold a position.
And this seems to be the main risk for Fact Machine: the project has to prove that its voting layer is broad and honest enough to function as an oracle, instead of becoming another playground for people who are better at coordinating than everyone else.
For crypto-native users, that kind of PvP may even be part of the fun. But for a new retail user, it is a different story. If they do not understand who is actually shaping the final vote split and whether they can trust it, it will be hard for them to deposit real money, even into the funniest opinion market on the internet.
Moni Score would slot in cleanly here. Show it next to a user's account and the crowd gets context it's currently missing: is this an established participant with real social presence, or a fresh profile with no reputation to lose?
During a market that shapes how you read the other side. After the vote reveals, it tells you whose read actually carried weight. For Fact Machine specifically, it's also a signal for catching multi-accounting before it distorts the outcome.
People will pay to be right first
Fact Machine is testing something very human. People love to argue, love being right, and love being right before anyone else catches on.
Prediction markets sell "I know what's going to happen." Opinion markets sell "I know how people think."
If the format catches on, it could end up broader than classic prediction markets. Clean, resolvable events are finite. Opinions aren't. Any news story, cultural fight, crypto narrative, or founder drama can turn into a market.
That's where the bet gets interesting. Win, and Fact Machine turns online disagreement into its own consumer category, a place where the internet doesn't just argue but trades on the argument. Lose, and it's a thin PvP game: the crowd formally settles the outcome, but nobody quite trusts who the crowd is. People show up once, lose money, and never come back.
Turning online arguments into a market
FactMachine is interesting not because it is another app on Solana.
It is interesting because it admits something obvious but underexplored: a huge part of the internet does not argue about truth. It argues about perception. What do people think is right? How will the majority interpret an event? Is this bullish? Is this cringe?
All of that already lives in comment sections, likes, quote tweets, polls, and endless arguments on X. Fact Machine is trying to turn that layer into its own market format: not just asking users what they think, but letting them trade on how they believe everyone else thinks.
The main question is still open: can Fact Machine make opinion markets broad, honest, and repeatable enough to become more than a fun alpha for people who enjoy guessing the mood of the crowd?
During the alpha testnet, the team partially validated the mechanics. Now the harder stage begins: will people regularly pay real money for their feel of the internet?