Market-structure
Prediction markets built for crowd wisdom are turning into speed contests as AI moves in, and a $2m pre-seed startup thinks natural language is the new order ticket
Elastics, founded by a former Goldman Sachs trader, is building an AI layer across Polymarket and Kalshi as automated bots already rank among the most profitable accounts on public wallets, arbitraging millisecond pricing gaps that used to be too small for manual traders to bother with.
Elastics raised $2 million in pre-seed funding to build what its founder Szymon Pawica calls an AI-native operating system for prediction markets, unifying access to Polymarket, Kalshi and Hyperliquid through a single interface with live and historical data, execution, and portfolio management in one place. The product it is building around, Trade with Words, is a natural language interface for deploying quantitative strategies without traditional order entry, letting a user describe a position rather than construct it manually. Pawica's framing is direct, as AI-driven automation becomes more widespread, manual trading is becoming increasingly challenging, which is less a prediction than a description of what is already happening on these venues.
The evidence for that is already visible on-chain. Public wallet analysis shows automated bots ranking among the most profitable accounts on these platforms, and arbitrageurs have extracted tens of millions of dollars from Polymarket by exploiting millisecond-level pricing gaps between venues, the kind of edge that only exists for participants fast enough to see and act on it before it closes. That is a meaningfully different market than the one prediction markets were originally built to be, where the value proposition was aggregating dispersed human judgment into a single price, not rewarding whoever has the fastest execution stack.
The infrastructure demands are shifting to match. Elastics and its competitors are building toward real-time data feeds, institutional-grade execution and latency management, the same category of plumbing that underpins traditional electronic derivatives markets, not the comparatively simple forecasting-platform stack prediction markets started with. The parallel the piece draws to FX markets, where machine-driven liquidity now dominates volume, is a reasonable one, except the shift into automation is arguably arriving faster here because these venues are younger, smaller and less structurally defended against it than FX was when algorithmic trading first took hold there.
What is worth watching is whether prediction markets can keep marketing themselves on the wisdom-of-the-crowd premise that built their retail user base once machine-driven liquidity is setting most of the actual prices. A market where a growing share of profit accrues to automated arbitrage rather than to participants with genuinely better forecasts is a different product than the one early users signed up for, even if the interface still looks the same. Whether that shift gets acknowledged openly or just quietly reshapes who wins on these platforms is the thing to track as volumes keep growing.
Read the original: Finance Magnates - Prediction markets built for crowd wisdom are turning into speed contests as AI moves in, and a $2m pre-seed startup thinks natural language is the new order ticket. Commentary is the independent editorial view of Share Trading; the original article is credited to its publisher.