Image credit: Source: LSports and gripAI product announcement. Never imply stock depicts the actual event.
Data provider LSports and insurtech firm gripAI launched Play Insurance in July 2026, a product that lets bettors pay a small, individually priced premium to protect a bet, then get paid back on a loss according to how close it came to winning. A losing near miss, the single most frustrating moment in sports betting, becomes a partial refund instead of nothing, and the sportsbook keeps the premium regardless of outcome.
The pricing is the product. gripAI's engine, trained on more than two billion historical bets, calculates each premium in real time using live odds, the number of selections in the bet, the player's own profile, and live match statistics. That is a materially different design from the flat rate cashback or bonus bet offers operators already run. A generic promotion prices risk the same way for every customer; Play Insurance prices it per bet, per player, in the moment the bet is placed.
In LSports' own pilot figures, 37% of players took the insurance offer when it appeared, and 39% of those who bought it once bought it again. Those repeat-purchase numbers are the detail that will get an operator's product team to take a meeting. A promotion that a third of players use, and that a similar share of users repeat, is behaving like a real product feature rather than a one-off acquisition gimmick.
Where the margin actually comes from
Cash out already turned the moment a bet is losing into a second revenue opportunity, and our analysis of cash out as a margin engine found operators quietly widening that spread well beyond fair value. Play Insurance works the opposite side of the same insight: instead of monetising a bet that is about to lose by buying it back cheap, it monetises the fear of losing before the outcome is known, charging a premium calibrated by an AI model that has seen two billion prior bets play out.
That makes it a natural companion to the products already reshaping sportsbook margins, same game parlays and micro-betting on individual match moments. Each of those products increases the number of discrete priced events inside a single bet slip. Play Insurance adds a second layer of pricing on top, a real time insurance premium calculated on a bet that has already been priced once by the sportsbook's own odds. For a supplier like LSports, already selling data feeds into the trading layer, bundling an insurance product on top of that same data is a logical extension rather than a new business line.
The dependency worth watching is on gripAI's pricing model itself. An insurance product only works if the premium genuinely reflects the probability of a near miss; price it too low and the sportsbook absorbs losses on every payout, price it too high and the 37% take-up rate collapses. That is exactly the kind of proprietary, patented pricing engine that becomes a supplier's real moat, the same dynamic that pushed Sportradar to hire Entain's Sameer Deen as suppliers race to own the layer of the business that used to sit inside the operator.
For bettors, the honest read is that Play Insurance is a product designed to keep them betting, not to change the underlying odds in their favour. Paying a premium for a chance at a partial refund is a form of risk management, similar in spirit to buying travel insurance, and like any insurance product the house has priced it to make money on average. It is worth using deliberately, as a way to manage a specific bet's downside, not as a habit that adds a second cost to every wager placed.
There is a responsible gambling question sitting underneath the product-design story. A tool that pays players back for losing, even partially, blurs the line between insurance and a disguised version of chasing losses. An operator rolling this out needs to be explicit that Play Insurance is a hedge against a specific outcome, priced like any insurance product to favour the house over time, not a way to reduce the overall cost of betting. Regulators that have spent the past two years tightening bonus and promotional rules, from the UK's advertising restrictions to the Netherlands' deposit controls, are likely to look closely at how insurance-style products are marketed once they scale beyond a pilot.
For smaller operators without LSports' data relationship or gripAI's pricing engine, the practical takeaway is that building an equivalent in-house is not realistic in the near term. Two billion historical bets is not a dataset a single-market sportsbook accumulates on its own, which means Play Insurance, like managed trading services before it, will likely spread through licensing rather than replication. The operators that adopt it first will be the ones already paying for premium data feeds, widening the gap between data-rich and data-poor sportsbooks rather than levelling it.
The near miss just became a product
Sportsbooks have spent years finding ways to keep a losing bet emotionally alive long enough to keep the player at the table. Play Insurance is the first version of that idea with its own priced, standalone revenue line, and if the repeat purchase numbers hold up at scale, expect every major supplier with a data feed to have a version of it within a year.