Cape Town, South Africa, May 5, 2026 — Prophet, an AI-native prediction market platform, has launched its first live trading tranche, introducing a system where an AI model acts as the counterparty to user trades using real capital.
The initial deployment allocates $10,000 in USDC to an AI-powered trading system and opens participation to users on the platform. Instead of matching buyers and sellers, the system allows users to trade directly against the AI, which generates probability-based pricing for each market.
What has launched
Prophet’s “Tranche 1” is a limited-access deployment designed to test the system under live market conditions. Users who deposit gain the ability to create markets, with the AI pricing each market upon creation. Once live, markets can be traded by other participants.
The AI system takes the opposing side of every trade, absorbing directional risk based on its probability estimates. Markets can resolve within relatively short timeframes, with some contracts settling in as little as 24 hours.
According to the team, the initial tranche is intended as a controlled test of system performance using real capital and user interaction.
One model, one probability
A key feature of the platform is its pricing mechanism. The system aggregates outputs from multiple large language models, including those developed by OpenAI, Anthropic, Google, xAI, DeepSeek, and Meta. These models independently evaluate each market question, with Prophet combining their outputs into a single probability estimate.
The same architecture is used for market resolution. When a market reaches its deadline, the system evaluates real-world outcomes and settles the contract without a formal dispute process.
The team notes that this approach is experimental and may be subject to limitations in interpretation or accuracy.
Why the prediction market industry is watching
Prediction markets have seen significant growth in recent years, though most platforms continue to rely on human counterparties and manual or committee-based resolution.
Prophet’s model introduces a different structure, where liquidity and settlement are managed programmatically. This may allow for faster market creation and resolution, though its effectiveness at scale remains to be assessed.
Trading as system feedback
The platform is designed to incorporate trading activity into its development cycle. Each trade generates data on pricing accuracy, while each market expands the range of scenarios the system must evaluate.
According to the team, this feedback loop is expected to inform improvements to the model in future tranches.
Risks and limitations
The current version operates without a formal dispute mechanism. Market outcomes are determined by AI-based interpretation, which may be subject to error.
The initial $10,000 allocation is limited relative to broader market standards, and the tranche is positioned as a testing phase rather than a full-scale deployment.
Regulatory considerations may also apply, as AI-driven prediction markets represent an emerging category with evolving oversight frameworks.
Next steps
Tranche 1 is scheduled to run through May 8, 2026. The team plans to use data from this phase to refine pricing, resolution, and system design ahead of future deployments.
Subsequent tranches are expected to expand capital allocation and user access.
About Prophet
Prophet is a machine that predicts the future. An AI with a bankroll trades directly against users, allowing for any market to be opened instantly. Prophet is solving liquidity and resolution for the long tail of prediction markets.
Platform: app.prophetmarket.ai
X: @prophetmarketai
Website: prophetmarket.ai
Media contact
Eneo Hollenbach
Chief Marketing Officer
team@prophetmarket.ai
Disclaimer
This press release is for informational purposes only and does not constitute financial advice. Participation in prediction markets involves risk, including potential loss of capital. AI-based systems may introduce additional uncertainties in pricing and resolution.
