TL;DR
It can enter, but through data distribution rather than liquidity. ICE has already opened Polymarket’s institutional gateway: a $2B investment at a $9B valuation in October 2025, followed by the launch of Polymarket signal and sentiment data products in February 2026. Commercial distribution of odds as a real-time information feed is already established.
A stress test using the flagship midterm-election market is less encouraging: for the same question of Democrats controlling the House, Kalshi’s open interest is about 4.8 times Polymarket’s and its 30-day volume about 4 times. Among Polymarket’s 500 trades over the latest 4 days, the top 10 account for 73.8%, while the median trade is only $20.6.
Pages and feeds can therefore generate continuing exposure, but price formation depends too heavily on a few large orders. This is where inadequate prediction credibility can undermine growth. Data is as of 2026-09-14 / 15; the price snapshot is dated 2026-09-19.
Platform level: usage exists, but Kalshi dominates turnover
Polymarket’s depth of committed capital is real; its depth of trading flow is not. As of 2026-09-14, platform-wide open interest was about $1.57B, rising steadily during September from $1.550B to $1.570B. Yet daily notional volume was only $52M–$78M. Kalshi’s open interest was $980M–$1.33B over the same period, but daily volume reached $1.6B–$2.56B.
| Metric (2026-09-14) | Polymarket | Kalshi |
|---|---|---|
| Open interest | $1.570B | $980M |
| Daily notional volume | $52.31M | $2.561B |
| Volume / OI turnover | About 0.03x | About 2.6x |
The conclusion is straightforward: Polymarket is a position-holding platform, where capital remains committed for a long time; Kalshi is a trading platform. Institutions need quotable prices refreshed through frequent turnover. Low turnover means slower odds updates and greater potential to lag news.
Midterm-election markets: the pages exist, but depth has not followed
Politics remains Polymarket’s anchor: political open interest is about $644M, or 41% of the platform, with daily volume of $2.1M–$3.2M. The flagship midterm markets, however, are too small for their headline role in the information feed.
| Market | Polymarket price | PM OI | PM 7-day volume | Kalshi price | Kalshi OI | Kalshi 7-day volume |
|---|---|---|---|---|---|---|
| Democrats control the House | 0.88 | $2.73M | $185K | 0.86 | $7.29M | $1.13M |
| Republicans control the House | 0.13 | $1.41M | $215K | 0.15 | $12.49M | $1.44M |
| Democrats control the Senate | 0.49 | $1.06M | $99K | - | - | - |
| Republicans control the Senate | 0.56 | $700K | $38K | - | - | - |
Over the last 90 days, daily volume in Polymarket’s market repeatedly ranged between $4,000 and $22,000, falling to about $4,050 per day in the latest 2 weeks. Kalshi’s corresponding market remained at $32K–$94K per day. Continuing use on Polymarket looks more like unclosed positions than daily repricing.
Another internal inconsistency deserves attention: the Senate prices are 0.49 + 0.56 = 1.05, while the House prices are 0.88 + 0.13 = 1.01. Pricing that deviates from 1 is typical of low-turnover markets. For institutions quoting odds as probabilities, this directly reduces credibility.
Credibility risk: a few trades determine the odds
The order book looks respectable on the surface: best bid/ask of 0.87 / 0.88, just a 1-cent spread, with about $430K of bids and $270K of asks within 5 cents. This can absorb directional trades of several hundred thousand dollars, but is far too small for institutional positions. A $2M order could move the probability several points.
The trade structure is more revealing. The 500 trades from 2026-09-11 to 09-15 totaled about $356K:
- The top 10 account for 73.8%;
- The 9 trades of at least $10K account for 71.6%;
- The median trade is just $20.6;
- Only 7 “smart money” wallets participated, with $38K bought and $19K sold, a net-long direction, and conviction of 1.4.
The price path also resembles slow drift rather than information-driven repricing: 0.80 in June, 0.89 at the end of August, and 0.87 now. If actual vote counts diverge significantly, media explanations will not say that market depth was inadequate; they will say prediction markets were wrong. That is the backlash mechanism.
Nonmarket risks are also tangible: South Korean police have tracked Polymarket users, with 26 arrested and 18 referred to prosecutors. The company also replaced 4 executives in a month. The first constrains organic growth in a key jurisdiction; the second weakens institutional confidence in product and compliance continuity.
Why data distribution is more realistic than institutional trading
ICE’s value is not immediately bringing large institutional orders into Polymarket. It packages odds as products for terminals, models, and research workflows. This bypasses the hardest current problem: institutions cannot directly deploy eight-figure capital into shallow on-chain markets, but can use their prices as supplementary signals.
The data is better suited to 3 workflows:
- Event-risk monitoring: tracking probability changes around elections, regulation, and policy;
- Sentiment and positioning proxies: cross-checking with polls, option skew, news sentiment, and fund flows;
- Anomaly alerts: triggering human review when prediction-market prices quickly diverge from traditional sources.
Adoption of a data product does not mean accepting its prices as true probabilities. To advance from quotable to calibrated, Polymarket still needs quality labels for depth, trade concentration, active market-maker counts, price impact, and cross-platform discrepancies. Otherwise users see a precise decimal quote without knowing that a few large orders may determine it.
What institutional information feeds require
To become a stable part of institutional workflows, Polymarket should not focus only on more markets and faster notifications. It needs a data-quality layer:
| Capability | Current issue | Institutional requirement |
|---|---|---|
| Probability prices | Quotes are visible, but liquidity quality is not adequately conveyed | Show spread, depth, impact cost, and trade concentration together |
| Historical series | Price changes can be observed | Annotate large orders, news events, rule changes, and abnormal trades |
| Cross-platform comparison | A single Polymarket price is easily misread | Standardize discrepancy monitoring against Kalshi, polls, and other benchmarks |
| Settlement reliability | Market rules determine final outcomes | Provide rule versions, dispute status, and settlement-source change logs |
| Compliance continuity | Users face enforcement risks in different jurisdictions | Clarify data licenses, access scope, and jurisdictional exposure |
The valuable product is not simply “Polymarket gives Democrats an 88% chance,” but:
Polymarket quotes 88% and Kalshi 86%; what is the depth, how many independent sources of capital moved the price in the last 24 hours, is the gap against poll models widening, and how sensitive is the price to $100K and $1M orders?
Only when the quality behind the price is exposed can prediction-market data advance from social screenshots to institutionally auditable signals.
Conclusion
The information feed is Polymarket’s best institutional gateway and its largest risk exposure. Through ICE terminal distribution, odds have reached institutional desks as signal and sentiment data. That route bypasses the practical limit on eight-figure on-chain orders, positioning Polymarket as a data supplier rather than a trading venue. Given current depth, it is the most realistic approach.
Yet if flagship election prices remain driven by a few daily large orders and Kalshi dominates volume by about 4 times, the gap between being widely quoted and being trusted will keep growing.
Core judgment. Using Polymarket’s midterm-election odds as sentiment and capital-positioning indicators is reasonable today; treating them as calibrated probabilities is riskier. Watch 3 things: whether flagship daily volume can remain above $30K and narrow the Kalshi gap; whether opposite-side prices converge consistently toward 1.00; and whether enforcement in South Korea and other jurisdictions spreads and suppresses organic active users. Without improvement, Polymarket can reach institutional screens but will struggle to enter core institutional pricing models.