Goldman bans its traders from betting on the next war
Goldman Sachs has told its trading desks they can no longer wager on the contracts that have come to define the new information economy: elections, inflation prints, war-and-peace. The move reads less like risk management than a confession.
On 10 July 2026, with a single internal memo, Goldman Sachs drew a line its competitors had not. The bank's employees can no longer trade prediction-market contracts tied to macroeconomic data and to geopolitics. The rule, first reported by Unusual Whales and confirmed by the bank's own compliance circular, sweeps in the two categories that have come to define the new information economy: who wins the next election, and when the next war ends.
The instinct inside the building is that this is a compliance story. It is not. It is a confession. A bank with a derivatives desk that can price a Brazilian real crisis to four decimal places has decided that the same employees cannot be trusted with a five-cent contract on whether the Federal Reserve cuts in September. The gap between those two skill sets is the actual news.
What the memo actually says
The prohibition covers contracts whose underlying events are macroeconomic prints and geopolitical outcomes. That is the language the bank itself is using, and it is the part of the rule that matters most. Macroeconomic prints are the bank's own bread and butter: payrolls, CPI, jobless claims, retail sales. The trading floor already has a privileged, often non-public read on those numbers, by way of client flow, primary-dealer positioning, and the quiet conversations that happen between sales desks and the Federal Reserve's primary dealers.
Geopolitics is a different problem. A Goldman analyst can publish a note on Taiwan, the Gulf, or the Russia–Ukraine front and move markets by lunch. The same analyst, on a weekend, can buy a Polymarket contract on the same outcome for the cost of a sandwich. The bank is now saying, in writing, that it cannot police the difference between those two acts. The line is a confession that the modern trading floor and the modern betting app have converged, and that nobody yet knows how to govern the seam.
The thin line between liquidity and leverage
The rivals are watching. Morgan Stanley and JPMorgan have not yet issued comparable guidance, according to the same reporting. That asymmetry is itself the story. A bank that bans its employees from a market while its competitors remain participants is not de-risking. It is buying optionality on a future regulatory shock, the kind that almost always arrives once the platforms grow large enough to threaten the incumbents.
The prediction-market complex is now a serious financial utility. Kalshi and Polymarket together processed hundreds of millions of contracts in the past year, on outcomes ranging from Federal Reserve decisions to ceasefires to cabinet shuffles in Brasília. Their pricing is now cited in broker research notes and, increasingly, by central banks looking for a real-time read on inflation expectations. The platforms have stopped being a casino side-bet and started being a price-discovery mechanism. The banks are late to that recognition.
What this means for the rest of the street
The compliance argument is real, and it deserves more airtime than it usually gets. Insider-trading rules were written for equities, bonds, and swaps. They were not written for an event contract that pays out on whether a ceasefire holds until Thursday. The legal definition of "material non-public information" in the prediction-market context is, charitably, unsettled. A trader who hears from a sovereign-desk contact that a talks track is alive has, in the strict legal sense, just heard something the market does not know. Whether buying a YES contract on the same outcome is a crime or a clever trade is the question every general counsel in lower Manhattan is now quietly asking.
But there is a sharper reading. Wall Street has spent fifteen years telling regulators that crypto is too small, too retail, too fringed to warrant the same supervisory architecture as the rest of the capital markets. The prediction-market platforms are on the same trajectory, only faster. Goldman's ban is a hedge: if the regulators come, Goldman wants to be the bank that said no first. If they do not, the bank has lost a few weekends of optional trading, which is the cheapest insurance policy in finance.
The stakes, plainly stated
The losers in this arrangement are the platforms, which now face a slower onboarding curve with institutional liquidity. The winners are the incumbents, who get a regulatory argument handed to them on a plate: that the new market is not yet ready for prime-brokerage treatment. The trader sitting at home with a phone, betting on whether the next CPI print comes in hot, is the most exposed of all: the platforms may now quote him tighter spreads, with thinner books, because the marginal professional liquidity is being asked to leave.
What remains genuinely uncertain is whether this is a Goldman-only story or the first move of a coordinated shift. The Wall Street firms have, in recent memory, copied each other's compliance memos within weeks once a category of conduct is flagged. The memo is dated 10 July 2026. Watch the other prime brokers through the rest of the quarter. The next name to publish a similar rule will tell us whether this was prudence, or positioning.
This piece sits at the intersection of market structure and information governance. Monexus framed the ban as a confession about the convergence of trading floors and prediction markets, rather than the wire-line read of a routine compliance update.