Two crypto prediction market startups said Monday morning that they were winding down, with both telling users to pull their money out by Sept. 30.
The two companies sat at opposite ends of the market. Trepa built its own mechanism on Solana and paid users based on how close their numerical guess came to the final outcome. Fireplace built a professional trading terminal on top of Polymarket and charged a fee for routing orders there. Neither found a durable place in a market now dominated by two exchanges.
According to DefiLlama data cited by The Defiant, Kalshi and Polymarket took 93.3% of the past 30 days’ volume across the 40 prediction market venues the platform tracks with volume adapters, or $13.87 billion of $14.87 billion. Kalshi alone accounted for 75.7%. The third-largest venue, BNB Chain-based OPINION, held 2.8%.
Funding levels show the gap just as clearly. Kalshi has raised $2.685 billion and was last valued at $22 billion. Intercontinental Exchange has invested $1.6 billion into Polymarket across two tranches. Trepa raised $420,000. Fireplace raised $1.5 million.
Trepa says its mechanism struggled to scale
Trepa’s founders, who signed their shutdown note as Jong and Leon, published roughly 1,200 words on what went wrong. The company raised its pre-seed in August 2025, led by Colosseum, with participation from Ignight Capital and angel investors including Balaji Srinivasan. It launched on Solana mainnet on Dec. 8, 2025. Its last playable round is set for Aug. 12.
The product used a pari-mutuel structure, splitting a pool according to how close each entrant’s number landed to the final result. That design required a crowd of users to predict within the same window. Trepa therefore ran one live hour a day, from 1 p.m. to 2 p.m. UTC, with 60-second rounds tied to bitcoin’s price.
“Retention degraded under that ask, and degraded more the longer someone had been with us,” the founders wrote. Expanding to more assets and more time windows would have thinned out the pools the mechanism depended on. “In the end we could not see a version of scaling Trepa that did not first make Trepa worse.”
The team also said both distribution channels it tried failed. “Referrals did not work and paid KOLs returned almost nothing,” they wrote. Conversations with partners about embedding Trepa’s widget kept circling back to the platform’s single daily hour, which existed because the mechanism needed concurrency. “Explaining your structural weakness as part of the pitch is not a strong position.”
Its founders also questioned the size of the addressable market
Trepa’s broadest claim went beyond its own balance sheet and focused on the market other startups are also chasing.
“Reported crypto DAU and MAU figures, including from sources we trusted, are softer than they look,” the founders wrote. “Our working conclusion is that the genuinely active audience for a product like this is in the hundreds of thousands at the moment, not the millions.”
They also pointed to geographic restrictions, which they said removed a large share of the addressable audience before the company even started, along with onramping friction. On the product itself, the founders reached a simple conclusion: “in consumer betting, simplicity may be the whole product.”
Retail trading outcomes support part of that view. Research covering 2.5 million wallets found that 84% of Polymarket traders are losing money, while only 2% have ever made more than $1,000 in profit.
Fireplace gave no reason for closing
Fireplace is operated by Hong Kong-registered Enclave HK Limited and was founded by Sumer Malhotra and Akshay Rajagopal. The company did not explain why it was shutting down. “It was a hell of a run,” the team wrote, adding that anyone building in prediction markets should get in touch if they wanted to talk about the technology.
Unlike Trepa, Fireplace did not run its own book. It aggregated third-party venues instead. Its terms of use list Kalshi and Polymarket as execution destinations, though its fee schedule mentions only Polymarket and charges a 1% taker fee “in addition to any fees levied by Polymarket.” Its wind-down page now sends departing users directly to Polymarket: “Fireplace runs on the same Polymarket markets, so your positions carry over.”
Fireplace launched publicly on Jan. 27, 2026, then announced its pre-seed on Feb. 17. The round was led by Frachtis, with White Star Capital and angel investors including Wintermute co-founder Evgeny Gaevoy. At the time, the company said its waitlist had topped 30,000.
Six months later, on July 13, Kalshi released Kalshi Pro, a free desktop terminal with a screener across roughly 2,000 markets and an integrated perpetuals interface. That put the market leader directly into the same product category Fireplace occupied, while already controlling about three quarters of total volume.
The front-end layer was already crowded
Even before Kalshi Pro arrived, the front-end wrapper layer was busy. DefiLlama tracks more than 20 interface wrappers routing to Polymarket, including Gate Predictions with $48.3 million in 30-day volume and Traderline with $61.3 million.
The wider backdrop is a weaker bitcoin price. According to CoinGecko, bitcoin was trading at $64,977, down 48.5% from its all-time high of $126,080 on Oct. 6, 2025.
That has not stopped capital from entering the category. ProphetX raised $35 million on July 28, and Pascal closed a $9 million Series A on July 16. Both are targeting the regulated U.S. market rather than trying to build a differentiated onchain mechanism.
Trepa’s founders said they still believe in the core idea
Trepa’s founders did not walk away from the underlying concept. “We still think there is something in the argument that people will pay to be measured on how close they get,” they wrote. “We were not the team that found the way to build a business around it.”