Updated Oct 4, 2026 9:48 PM ET.

For investors

iPredictSport is privately owned, growing fast and open to investment. Metrics as of October 2026.

11

sports covered

Tennis, NFL, college football, golf, NBA, college basketball, NHL, UFC, MLB, soccer and chess.

632,000+

pro tennis results archived

ATP, WTA, Challenger and ITF since 2020, timestamped and append-only.

33,000

settled markets with price paths

NFL, college football, golf and UFC exchange markets, hour by hour, to settlement.

1,900+

production commits in 12 weeks

Shipped since July 2026. New sports go from raw prices to a live public board in days.

Why now

A new market is opening

Sports prediction markets on regulated exchanges like Kalshi, and on Polymarket, have gone from a niche to a mainstream way to trade sports. The traders arriving there want numbers, tools and a crowd. That's what we build.

Trust is the scarce asset

Sports betting media is full of touts selling "locks". We publish every call before the game and grade it against the market price, misses included. In a category short on credibility, that record is the moat.

AI agents need data

Our predictions and price history already serve over JSON and MCP, so AI agents and trading bots can use them directly. We're a data source for the agent era, not just a website.

What's already built

  • Measured model edge where it exists. On the WTA, our model plus market-reading layer beats the exchange price six hours before the match by 9.5 milli-Brier on 253 out-of-sample matches, with the confidence interval clear of zero. The women's model alone is at parity with a bookmaker on 4,077 matches and ahead of it on grass. Where we don't beat the price yet, we say so in public.
  • A data asset that can't be bought after the fact. Point-in-time prices, order books and our own predictions, each pinned to the model version that made it. That includes 9,727 tennis derivative markets with 285,571 raw price candles. A missed observation can't be recovered later, so this archive only gets harder to copy.
  • One engine, many front-ends. A shared sports engine feeds every board, the JSON and MCP feeds, the paper-trading agent and Greta, our AI sports host. Adding a sport means a new pipeline on shared code, not a rebuild.
  • Automated operations. Supervised collectors, scheduled publishing and checks that fail loudly when a number drifts. A very small team runs eleven sports.

How we make money

Partners and advertising

Referral revenue from exchanges and sportsbooks, plus display advertising, on a free site built for search traffic. Live now.

Paid feed and API

Fresher predictions, more fields and keyed API access, tiered by delay. Built on the feed that already powers the site.

Data licensing

Historical point-in-time market and prediction data for funds, platforms and researchers. B2B contracts on an archive nobody else kept.

The IPS Crew

A community of traders, analysts and bot builders, with Discord, pick'em leagues and a bot leaderboard. It drives retention, referrals and premium tiers. See the crew.

Where the money goes

  • People. We're hiring a Senior Developer and a VP of Marketing to build the community.
  • Growth. Content, creators and the crew, plus paid acquisition tests with kill rules.
  • Product. The paid feed and API, more sports, and the trading-agent tooling.
  • Data. More markets and more depth, collected now, while the window is open.

Talk to us

Want the deck, the data room or a walkthrough of the track record? Alex, the founder, takes every investor conversation personally. Read his background.

Request investor materials

This page is for information only. It is not an offer to sell, or a solicitation of an offer to buy, any security. Any offering will be made only through definitive documents, and only to investors who qualify under applicable securities laws. Model results described here are measured on historical and paper data. Past performance does not guarantee future results.

Tennis board Contact