Woofun AI reports that Goldman Sachs has instituted a comprehensive ban on employee participation in prediction markets, marking a sharp reversal from CEO David Solomon’s previous characterization of the sector as "very interesting." This policy shift effectively closes a door that was only opened six months ago, reflecting a rapid hardening of corporate stances toward decentralized betting platforms. The prohibition extends beyond simple speculation, targeting contracts tied to corporate restructuring, election outcomes, and macroeconomic indicators, thereby signaling a significant tightening of internal compliance frameworks across major financial institutions.

The updated internal personal trading policies, implemented on July 9, explicitly forbid employees from engaging in contracts related to specific company events, including whether Goldman Sachs itself would initiate an acquisition or restructure within a given quarter. The restrictions also cover election results, financial market performance—specifically including Bitcoin prices—macroeconomic data releases, geopolitical developments such as ceasefire timelines in active conflicts, and the regulatory outcomes of pending mergers and acquisitions. While betting on sports and entertainment remains permissible, the scope of prohibited activities is broad. Penalties for non-compliance are severe: employees who violate the rules twice face dismissal or account closure.

Furthermore, if a transaction is deemed improper, the firm reserves the right to recover profits exceeding $200 or donate that amount to charity. A Goldman Sachs spokesperson declined to comment on specific policy details, reiterating only that trading using material non-public information (MNPI) is prohibited in all markets where the firm operates.

Notably, the approach to prediction markets varies significantly across Wall Street firms, revealing a fragmented regulatory landscape within the financial sector. Point72 Asset Management and Balyasny Asset Management have adopted more drastic measures than Goldman Sachs, implementing complete bans on prediction market trading for all individual accounts without any exemptions. In contrast, JPMorgan has merely advised employees to "be cautious" without issuing an official prohibition. Bank of America is currently pushing new trading restrictions to its workforce, while Morgan Stanley has included relevant provisions in its employee code of conduct. These divergent strategies highlight the uncertainty surrounding the legal and ethical boundaries of prediction market participation among financial professionals.

Structurally, Google’s Chrome app store has also moved to cut off distribution channels for these platforms. The updated developer program policies, which will take effect on August 1, 2026, explicitly list prediction markets as prohibited categories. This update forbids extensions that facilitate or support real-money transactions based on prediction results.

Additionally, the new policy mandates that user data collected by extensions be used solely for the stated purpose, prohibiting any other use, and bans extensions designed to circumvent AI-driven security measures. Google advises developers to review the compliance of their existing extensions as soon as possible. After August 1, if an extension is found to violate relevant policies, the Chrome app store may take enforcement action.

The impact of these restrictions on platform access and user data is substantial, particularly for platforms like Polymarket and Kalshi. While Google’s ban does not affect the web and mobile versions of these platforms directly, Chrome extensions serve as a critical access point for some users. Blocking this channel is equivalent to setting up a barrier at the browser level, potentially reducing user engagement and accessibility. The policy change also includes stricter requirements for transparency in user data collection. Data collected by extensions can only be used for the declared purpose, and any changes in data handling must be notified to users proactively. This heightened scrutiny on data usage adds another layer of complexity for platforms relying on browser-based integrations.

Woofun AI data shows a more critical variable is the enforcement case involving Michele Spagnuolo, which served as the direct trigger for Goldman Sachs’ ban. In May this year, the CFTC announced that Google software engineer Michele Spagnuolo was suspected of making $1.2 million by illegally trading prediction contracts on Polymarket. He was charged with fraud and money laundering. The contract in question was related to "who will have the highest search volume in 2025." In its lawsuit, the CFTC sought compensation, restitution, civil fines, trading and registration bans, as well as a permanent ban for further violations of the Commodity Exchange Act and CFTC regulations. Yesterday, the U.S. Attorney’s Office for the Southern District of New York announced criminal charges against Michele Spagnuolo in the same court.

The details of the illegal betting activity reveal a sophisticated exploitation of internal data. Last year, Michele Spagnuolo used Google Year in Search data, accessible only to a few employees, to place bets on the most popular search figures of the year on Polymarket through the account "AlphaRaccoon." This involved 25 transactions. He accurately predicted that Kendrick Lamar and d4vd would make it into the top five. This is the first insider trading case involving a private company in prediction market history. Previously, discussions about insider trading in prediction markets mainly focused on political events such as elections. The Spagnuolo case proved for the first time that corporate employees could use internal company information to profit from prediction markets. For Wall Street firms that come into contact with large amounts of non-public financial information in their daily operations, such risks are particularly prominent. If a trader knows about upcoming earnings data, M&A plans, or regulatory decisions and places bets on related contracts on Polymarket or Kalshi, it is no different in nature from insider trading in traditional securities markets.

However, the current regulatory framework and identity verification mechanisms for prediction markets are far less mature than those in the securities market.

Regulatory investigations and legal challenges are intensifying against the backdrop of these corporate actions. On June 26, the CFTC announced a thorough investigation into Polymarket, involving fake transaction videos, false winning records, and undisclosed paid promotion activities. On the same day, consumer rights organizations sued Polymarket, along with its CEO and CMO, in Washington, D.C. The CFTC has filed federal lawsuits in nine states to claim jurisdiction over prediction markets, while 17 Democratic senators are trying to prevent the CFTC from using federal funds to sue state governments. This legal tug-of-war highlights the jurisdictional ambiguity surrounding prediction markets, with federal and state authorities vying for control.

Global restrictions and market growth metrics present a complex picture of the industry’s trajectory. Argentina ordered a nationwide blockade of Polymarket in March this year, becoming one of more than 30 countries around the world that have imposed access restrictions on prediction markets. This ruling also forced Google and Apple to remove Polymarket’s apps from their app stores in Argentina. Despite these restrictions, the market continues to grow. As of June 22, the monthly nominal trading volume of prediction markets reached a record high of $291.38 billion.

Moreover, Kalshi is seeking a new round of financing at a valuation of $40 billion, and ICE, the parent company of the NYSE, has invested $2 billion in Polymarket, with capital continuing to flow in. This influx of capital suggests strong investor confidence, even as regulatory headwinds increase.

The definition debate and future outlook remain uncertain. Selig, the head of the U.S. CFTC, insists that prediction markets are federally regulated financial derivatives, and for this reason, he has sued nine states that attempt to regulate prediction markets under gambling laws. An analysis by the Wall Street Journal shows that over 70% of accounts on Polymarket are in the red, with only 0.1% of accounts accounting for 67% of all profits—a pattern more typical of casinos. But it is clear that restrictions are coming from multiple directions simultaneously: federal law enforcement investigations, political pressure from Congress, state governments’ attempts to assert jurisdiction under gambling laws, internal compliance restrictions from Wall Street firms, and blockages of distribution channels by tech platforms—are all accelerating the encirclement of prediction markets. This marks a pivotal moment for the industry, as the gray area surrounding prediction markets narrows by the day.