Key Numbers

  • 22 billion executed retail trades used to train BayesShield AI
  • Users can earn profits without risking their own capital
  • Deposits of $1+ unlock higher payouts, held in US Treasury Bills
  • Platform launched Tuesday by former FTX Europe exec Patrick Gruhn

Bottom Line

UpsideOnly flips the retail trading model by letting users provide ideas while the firm trades with its own balance sheet. The AI, trained on a massive retail dataset, decides which ideas to act on, freeing users from loss risk while offering profit participation.

Retail investors can now test market insight on a platform that monetizes crowd wisdom and machine learning, potentially reshaping how retail participation is compensated.

UpsideOnly’s launch on Tuesday marks the first time a platform has combined a patent‑pending AI model trained on 22 billion retail trades with a payout mechanism that rewards users for successful signals while shielding them from downside. Former FTX Europe executive Patrick Gruhn announced the platform, stating that it “completely changes this broken structure” by aligning human insight with firm capital.

AI‑Driven Prediction Engine Trains on Unprecedented Retail Data

BayesShield AI, the core of UpsideOnly, is a structural departure from traditional retail trading. The model ingests more than 22 billion executed retail trades, the largest behavioral trading dataset ever assembled for training. By merging crowd‑sourced predictions with machine‑scale pattern recognition, the algorithm refines its signals in real time as users submit directional bets on equities, crypto, commodities, and forex.

When the AI’s confidence exceeds predefined thresholds, the firm’s balance sheet executes the position. If the trade fails or is not placed, users bear no financial consequences. This design eliminates the typical 12 billion‑dollar annual loss suffered by retail traders, according to the team.

Zero Capital, Zero Risk – Users Earn from Company Trades

UpsideOnly removes the need for users to deposit margin or capital. Instead, participants can earn from the company’s own trades. The platform offers higher payouts to users who place a refundable deposit of $1 or more. The deposit, held in US Treasury Bills in an external fiduciary account, is never used as a trading stake and can be withdrawn at any time.

By decoupling user liability from trade execution, UpsideOnly promises a “win‑together” model where human insight fuels the firm’s capital, aligning incentives for both sides.

Why This Matters

For retail investors, UpsideOnly offers a risk‑free way to monetize market insight. The platform’s AI leverages a vast retail dataset, potentially generating higher‑quality signals than conventional crowd‑source or algorithmic models alone. If the model performs as projected, it could redefine profit participation and reduce the systemic loss burden on retail traders.

What to Watch

  • Watch: UpsideOnly’s user adoption metrics in the first quarter post‑launch (expected Q3 2026).
  • Next catalyst: Performance of BayesShield AI against benchmark indices (reported quarterly).
  • Watch: Any regulatory filings from UpsideOnly regarding the use of user data for AI training.
  • Next catalyst: Public disclosure of the company’s balance sheet exposure to user‑generated trades.