On any given Sunday in the fall, somewhere between 30 and 40 million Americans place a bet on an NFL game, and most of them are losing to a sportsbook whose only edge is a better spreadsheet. Rithmm, a Boston startup founded in 2022, is selling those bettors a spreadsheet of their own. The product lets a user build a full predictive model, drag a few sliders to weight the factors they care about, and then watch the model issue an opinion on every matchup of the week [Rithmm website]. It is, in spirit, a Bloomberg Terminal for the guy at the bar who insists he has a system.
The company launched publicly in March 2023 with a seed round led by Boston Seed Capital, and a cap table that includes DraftKings, Accomplice, Counterview Capital, Permit Ventures, Correlation Ventures, Oyster Ventures, and Service Provider Capital [BusinessWire, March 2023]. Total disclosed funding has since grown past $3 million [SBC Americas, March 2025]. The presence of DraftKings on that list is the tell: the largest US sportsbook is not in the habit of writing checks to companies it considers a threat, which suggests Rithmm is being read as a top-of-funnel tool that keeps engaged bettors engaged, rather than an arbitrage weapon pointed at the house.
The wedge is the custom model. Existing tools in the category, Action Network chief among them, sell content, picks, and odds comparison. Rithmm sells the underlying machinery: simulate every game, every player, every sport, then track how the model's predictions actually performed [Rithmm website]. Free AI player props sit at the top of the funnel [Rithmm website], and Core and Premium subscription tiers sit behind them, with NFL and college football products launched ahead of the 2023 season [PRNewswire]. The pricing page carries the standard entertainment-only disclaimer [Rithmm website], which is both a regulatory shield and a quiet acknowledgment that the unit economics here run on subscriptions, not on a cut of action.
Why the bet has shape
The legalized US sports betting market has expanded to roughly 38 states since the 2018 Supreme Court ruling that struck down PASPA, and the bettor population is now large enough to support a tools layer above the books themselves. That tools layer is what Action Network proved exists, what Outlier and Pikkit are scaling into, and what Rithmm is differentiating inside by handing the model itself to the user. The interesting structural fact is that books like DraftKings and FanDuel benefit from bettors who feel sophisticated, because sophisticated-feeling bettors place more bets. A model-builder hits that emotional note exactly. It also explains DraftKings sitting on the cap table without obvious channel conflict.
The team is small but credentialed. CEO and co-founder Megan Lanham leads the company [Crunchbase], which traces its origins to MIT [Gaming Americas, August 2024]. Senior manager Brian Beachkofski holds an MIT Sloan degree [LinkedIn], and the engineering and product bench includes Shirshak Sharma and Dylan Hoose [LinkedIn]. The most strategically interesting hire is on the advisory side: Paris Smith, the former CEO of Pinnacle, one of the sharpest sportsbooks in the world, joined as an advisor and made a follow-on investment [Gaming Americas, August 2024]. Pinnacle built its reputation on accepting sharp action that other books refuse, which means Smith knows precisely what a model-driven bettor looks like and what they will pay for. On the content side, analyst Josh Appelbaum anchors shows including CBB Live and Sunday Sweats [Rithmm YouTube; Rithmm website], giving the product a media surface that is cheaper than paid acquisition.
The funding picture
| Round | Date | Amount | Lead |
|---|---|---|---|
| Seed | March 2023 | undisclosed | not disclosed |
| Seed | May 2023 | $2.0M | Boston Seed Capital |
| Cumulative through 2025 | March 2025 | >$3.0M | n/a |
Sources: [BusinessWire, March 2023]; [Tracxn, May 2023]; [SBC Americas, March 2025].
The honest counterfactual
What bears will say is that Action Network already owns the bettor-tools category in North America, with years of SEO, a podcast network, and an exit to Better Collective behind it, and that a custom-model product is a feature a larger competitor can copy in a quarter. They will also point to the entertainment-only positioning as a ceiling on willingness to pay. What bulls answer is that Action Network's product is fundamentally editorial: it tells bettors what professional analysts think. Rithmm's product is fundamentally generative: it lets bettors build and back-test their own thesis, which is a different psychological transaction and one that the incumbent has not chosen to build. The Paris Smith follow-on is the strongest external signal that someone who has run a real book believes the second product has a real customer [Gaming Americas, August 2024].
Back of envelope
Assume a Premium subscription in the $30 to $50 per month range, consistent with category pricing, and assume Rithmm needs to clear roughly $5M ARR to justify a Series A at a typical seed-to-A multiple. At $40 per month per user, that is about 10,400 paying subscribers. The US sports betting handle in 2024 was reported at roughly $150B, which implies tens of millions of active bettors. Rithmm needs to convert about one in every two thousand of them. That is a marketing problem, not a market problem, and it is the right kind of problem to have.
What to watch
The next twelve months will be defined by the 2025 NFL and college football seasons, which together account for the majority of US bettor engagement and the company's clearest revenue window. Watch for a Series A announcement, an expansion of the player props product into NBA and MLB at full depth, and any sign that DraftKings or another book formalizes the relationship beyond a passive equity stake. The backend developer role currently posted on the careers page [Rithmm career page] suggests the team is still building, not yet scaling.
The company Rithmm has to beat is Action Network. Action owns the eyeballs and the editorial voice. Rithmm is betting that bettors would rather build the model than read about someone else's. If that bet is right, the editorial voice becomes the slower business.