On 13 August 2026 two complaints were docketed in the Circuit Court for Baltimore City, and the name at the top of both is worth reading slowly. The plaintiff is not the State of Maryland and not its gaming regulator. It is City of Baltimore, ex rel. Ebony Thompson, suing through its City Solicitor under an ordinance the city wrote for itself. The complaint against KalshiEX LLC and six other defendants carries case number C-24-CV-26-005532; the complaint against the three Polymarket entities is C-24-CV-26-005535. They were filed twelve minutes apart. For anyone drawing the jurisdictional map an event contract must survive, this is a layer that was not on it.
The caption names a city solicitor, not a state attorney general
Every prediction market exchange has a mental model of who can come after it: the Commodity Futures Trading Commission federally, and a state attorney general or gaming commission at the state level. The Baltimore filings fit neither slot. Both open with the City of Baltimore suing "by and through its City Solicitor, Ebony Thompson of the Baltimore City Law Department." The signature block adds Sara Gross, Chief of the Law Department's Affirmative Litigation Division, and Chief Solicitor Thomas P. G. Webb.
Beside them appear four lawyers from DiCello Levitt LLP, pro hac vice motions still to be filed. The Maryland Daily Record reported on the day of filing that the same firm already represents Baltimore against DraftKings and FanDuel, against sweepstakes casino operators, against Philip Morris International, and over the ship that struck the Francis Scott Key Bridge. That is a forecast rather than gossip: this layer is not staffed by one city's in-house lawyers alone, but by a repeat-player firm that can carry the theory to the next city cheaply.
The cause of action belongs to the city and the definition is borrowed
The architecture is a two-step, and the second step gets missed. The claim arises under the Baltimore Consumer Protection Ordinance, which bars "unfair, abusive, or deceptive trade practices" in connection with consumer goods and services at Baltimore City Code Article 2, section 4-2 and imports its definitions from the Maryland Consumer Protection Act at section 4-1. Those definitions, at Maryland Code, Commercial Law section 13-301, reach misleading representations, claims of an approval a merchant does not have, and failures to state a material fact where the omission tends to deceive.
Gambling law never supplies the cause of action, only the predicate. The complaints argue the contracts meet the state definition of sports wagering at Maryland Code, State Government section 9-1E-01(j), which covers "the business of accepting wagers on any sporting event by any system or method of wagering" and then names eleven methods, exchange wagering among them. Once that predicate is assumed, representing the product as lawful in Maryland becomes the deceptive act and operating without the state licence becomes the unfair one.
The consequence is structural. An exchange that wins the preemption argument about state gambling enforcement has not won against a consumer protection ordinance, because the ordinance is not a gambling law. It is a truthfulness law reading a gambling statute to decide what was true.
A promise not to enforce has a party list
Maryland's own enforcement against Kalshi has been on hold for over a year. After the federal district court declined to enjoin the Maryland Lottery and Gaming Control Commission, Kalshi told the court that on 12 August 2025 it had received written assurances from the Commission's counsel that the state would forgo enforcement of the relevant Maryland laws pending its appeal, making an injunction unnecessary. Kalshi withdrew the motion. The Fourth Circuit heard argument on 7 May 2026 and, as of 9 October 2026, has not ruled.
So the state had stood down. Almost exactly one year later the city sued anyway. Nothing improper happened: Baltimore was never a party to the case where that assurance was given, and a promise by one government's counsel does not bind a different government that was not in the room. We have written before about how pauses negotiated with a state regulator vary by who negotiated them. Baltimore is the sharper version: a pause is only as wide as its party list, and the party list of a preemption suit against a state commission does not include that state's municipalities.
A venue treating a standstill agreement as peace in a jurisdiction has mistaken a contract for a condition.
Eight counts covered one exchange and three places to reach it
The two complaints are built differently, and the difference maps distribution. The Polymarket complaint runs two counts, deceptive and unfair practices, against QCX LLC, Blockratize Inc. and QC Tech LLC. The Kalshi complaint runs eight, the same two theories under four headings: claims against Kalshi, Robinhood, Webull and Coinbase.
The defendants are KalshiEX LLC and Kalshi Inc., Robinhood Markets, Inc. and Robinhood Derivatives LLC, Webull Corporation and Webull Financial LLC, and Coinbase Financial Markets, Inc. Three of those four families neither list contracts nor resolve them. They offered the surface a Baltimore resident touched, and the city treated that surface as a distinct violation by a distinct merchant, with its own count and penalty exposure.
This is the practical content of the idea that the first screen a trader sees belongs to someone else. A perimeter drawn around the exchange alone leaves three other companies holding identical claims, none of which can be answered by pointing at the exchange's designation.
What the distributors were not accused of saying
The counts are not copy-paste, and the gap is instructive. Against Kalshi, the deceptive practices count alleges two kinds of misrepresentation: that the wagers are lawful in Maryland, and that the markets are trustworthy, transparent, safe and protected against insider trading. It adds a failure-to-disclose theory about how far ordinary users trade against participants holding informational, technological or financial advantages.
Against Robinhood, the equivalent count keeps only the legality representation. The integrity claims drop out. That is not generosity; the city is matching each allegation to whoever made the statement. Representations about market integrity attach to the party that published them.
Read as a compliance signal rather than litigation risk, that is nearly a drafting rule. Claims about safety, transparency and insider trading protection have to be evidenced, because they travel with the publisher and are actionable separately from the licensing question. Directive 02 asks that conflicts and settlement sources be stated plainly on the surface where the trade happens. A complaint that separates the integrity promise from the legality promise shows where an unevidenced one lands.
The penalty clock runs by the day and by the violation
What makes a municipal ordinance more than a nuisance is in Baltimore City Code Article 2, section 4-3. The civil penalty is capped at $1,000, but subsection (b) makes each violation a separate offence and subsection (c) makes each day a violation continues a separate offence. A thousand dollars multiplied by accounts and days is no rounding error, and the complaints ask for the maximum. Whether that survives a court is another question; the point is that exposure is computed from a period of operation, so the clock ran before anyone at the exchange read the ordinance.
Two remedies the ordinance does not name
The demand for relief asks for statutory penalties, an injunction barring the defendants from accepting transactions from Baltimore residents, disgorgement of gains and restitution to consumers. The provision the city sues under, section 4-5, is narrower: subsection (d) lets the City Solicitor seek injunctive relief and the imposition and collection of civil penalties, in addition to any other enforcement action authorised by law. Disgorgement and restitution are not named.
That residual clause may well carry them, and we will not predict how a Maryland judge reads it. What is observable is that the two largest money remedies in the prayer rest on a phrase rather than an express grant. A smaller sign points the same way: all ten counts across the two complaints claim their civil penalty under section 4-4, but section 4-4 is the criminal penalties provision, and the civil penalty they describe lives in section 4-3, which the demand for relief cites correctly. Nothing of substance turns on it; both sections carry the same figure and per-day structure. It matters as a reminder: this is a body of law not yet litigated often enough for its pin cites to be automatic. An exchange cannot extrapolate a municipal claim from state ones, because the people drafting them are also doing this for the first time.
Removal answered the venue and not the claim
The exchange's response was procedural. A prediction market litigation tracker maintained by Mick Bransfield records that Kalshi removed Baltimore's suit to federal court on 14 August 2026, one day after filing, and lists the federal case as 1:26-cv-03217 in the District of Maryland before Judge James K. Bredar. We could not reach the docket, so that comes from the tracker, not the filings. The tracker also lists remands to state court in several other Kalshi removals, so the venue question is itself unsettled.
Kalshi's public position was a merits position. The company told the Maryland Daily Record the suits were "clearly political theater by Mayor (Brandon) Scott" and that it operates lawfully under federal jurisdiction. In the city's own announcement, City Solicitor Ebony M. Thompson said the companies "cannot circumvent Baltimore's consumer protections by repackaging gambling as something else." Neither statement resolves anything, because removal decides which courthouse hears the ordinance claim, not whether the ordinance reaches the conduct.
Our reading of preemption has not changed since we wrote that the one sentence of the Commodity Exchange Act naming gaming describes places a designated contract market is not. The complaints make the mirror-image argument, leaning on the congressional statement at 15 U.S.C. section 3001(a)(1) that states should have primary responsibility for gambling inside their borders, and on Ah Sin v. Wittman, 198 U.S. 500 (1905), for suppressing gambling as a core police power.
Every harm allegation is a number somebody could publish
Strip the legal theory away and the factual core of both complaints is a set of measurements, each something a venue could report about itself.
- The age floor. Maryland prohibits sports wagering by anyone under 21 under COMAR 36.10.13.44, a one-sentence regulation. Both complaints allege, citing Kalshi's own published answers as visited on 27 July 2026, that eighteen-year-olds could trade sports contracts on the platform. The Maryland Daily Record reported the same gap. It is the cleanest allegation in either document, because it compares two published numbers.
- Who loses. The complaints cite an analysis by Brad Lipton and Toyosi Odusola published by the Roosevelt Institute on 7 July 2026, putting net retail losses on Kalshi at $583.5 million through 15 May 2026, more than two thirds of it from sports. Kalshi disputes it, issuing a rebuttal on 10 July 2026 asserting significant flaws in the methodology. We have not read that rebuttal in full, so both stand as positions.
- The ratio. The Kalshi complaint asserts at paragraph 107 that the company itself acknowledged nearly three times as many people losing money as making it, citing a Wall Street Journal report of 3 May 2026 by Neil Mehta, Katherine Long and Caitlin Ostroff. That report is paywalled and we have not read it, so this is the complaint's characterisation of it.
Notice what the list has in common. The city's harm case rests on a contested outside estimate and on a figure attributed to the company through a paywalled newspaper, not because the city was lazy but because the category publishes no agreed figure of its own. We have argued that nobody counts who stayed. These complaints are what happens next: when a venue does not publish the number, somebody else's estimate becomes the number of record, and the venue argues methodology from behind.
Directive 06 says revenue should come from many informed participants over years rather than a few ruined ones over months. That is unfalsifiable until a venue publishes the ratio of accounts finishing a period down against those finishing up. Publishing it quarterly would concede nothing the Roosevelt Institute has not already asserted, and would own the denominator.
The map had three layers and now it has a fourth
We wrote two weeks ago that an exchange draws a federal layer and a state layer, and that a third arrives when the buyer stands on tribal land, where the unit is a parcel rather than a state. Baltimore supplies a fourth with a different shape again. A municipal consumer protection ordinance is not a smaller version of a state gambling statute. It has its own enacting body, enforcement officer, penalty clock and litigation counsel, and it is bound by nothing its state agreed to.
Baltimore's ordinance was broadened only in 2023 to reach harms against residents rather than to the city itself, and most Maryland municipalities have nothing comparable. But it exists, the theory is drafted, and the firm that drafted it works for more than one city. Directive 05 asks that standards be set against the jurisdiction that regulates next, not only the one that has not yet. That jurisdiction may not be one anybody at the exchange has a name for.
What holds whichever way the Fourth Circuit rules
The Fourth Circuit could hand Kalshi a clean preemption win and the ordinance claim would still need an answer, because it asks whether statements were true, not only whether a licence was required. So: what can a venue do that does not depend on the result?
Map below the state line. A jurisdiction list whose smallest unit is a state cannot say whether a resident sits inside a municipality with its own ordinance and solicitor. That is a resolvable data problem, and the tribal-lands analysis already showed parcel-level answers are buildable.
Separate the two kinds of claim. Legality statements and integrity statements carry different exposure and attach to different parties. Whoever publishes "protected against insider trading" should be able to produce the surveillance record behind it. The CFTC's Director of Enforcement, David I. Miller, addressed insider trading in prediction markets at NYU on 31 March 2026, a speech the complaint cites. An integrity claim is checkable against a public enforcement agenda.
Treat distribution as inside the perimeter. Three of the four defendant families in the Kalshi complaint are venues through which the contract was reached. If a broker's screen makes the legality representation, the broker owns a count.
Publish the denominator. The age floor, the loss ratio and the concentration of profit are the three facts these complaints are built from, and all three are measurable in-house. Reporting them on a schedule hands a plaintiff no ammunition it lacks. It replaces a contested outside estimate with its own audited number, which is the only version of that argument it can win.