Government and Policy Sector

Kalshi is blocked from operating in Washington state, Michigan, and Nevada as the prediction market platform battles state regulators in court while its federal overseer, the CFTC, drafts new rules to assert jurisdiction over the industry. A Washington state court issued a preliminary injunction last week barring Kalshi from offering contracts on sports, elections, politics, culture, technology, science, and other events, finding the company likely violated state gambling and consumer protection laws. Kalshi notified customers this week and filed a motion asking the judge to reconsider, arguing that Washington state has agreed not to enforce the same restrictions against Crypto.com, a competitor offering identical event contracts. The legal fight is escalating across multiple fronts: Connecticut and New York have sued Kalshi in federal court, with New York seeking at least $36 billion in damages and calling it an unlicensed gambling operation; the Second Circuit denied Kalshi's request for temporary relief in Connecticut; and the CFTC has issued an emergency order allowing Kalshi to continue trading even if a court orders it to stop. At a CFTC roundtable on August 21, CME Group Chairman Terry Duffy called prediction market operators "carnival barkers" and questioned whether they face the same regulatory scrutiny as established exchanges, while Kalshi co-founder Luana Lopes Lara shot back that CME should "learn a bit about efficiency." The CFTC, which claims exclusive federal authority over prediction markets, is simultaneously drafting new rules on consumer protection and product governance, a move critics say is an attempt to shore up the agency's legal position after courts have sided with states. A U.S. survey published August 12 found that 79% of prediction market users lost money in the past year, with 51% using borrowed funds, adding a consumer protection dimension to a dispute framed primarily as a jurisdictional battle between federal and state regulators.

President Trump announced Friday that the U.S. will allow 300,000 metric tons of ground beef to enter duty-free for 90 days, waiving the out-of-quota tariff that normally reaches 26.4% on imports above country-specific limits. Trump said he has secured commitments from foreign exporters to sell the beef at 25 percent below current market prices, which averaged $6.89 per pound in July—up 57% from five years ago and near record highs. The move reverses Trump's longstanding tariff-first logic: with the U.S. cattle herd at multi-decade lows due to drought and high feed costs, domestic supply cannot meet demand, so tariffs on imports simply raise consumer prices rather than protect producers. The administration also plans to lift a yearlong ban on cattle imports from Mexico within two weeks via executive order. The National Cattlemen's Beef Association immediately objected, arguing that flooding the market with government-subsidized foreign beef discourages domestic herd rebuilding and sacrifices long-term stability for short-term political messaging ahead of November midterm elections. Tyson Foods, which closed two beef plants and cut over 2,500 jobs, has said prices are unlikely to recover until 2027 at the earliest. The White House did not name the exporters making price commitments or specify which countries are part of the deal.

The FDA published a discussion paper Tuesday seeking public input on how to regulate generative AI medical devices, proposing a "competency-based" premarket evaluation modeled on physician licensing exams. The agency's Center for Devices and Radiological Health would test the final user-facing AI product—not underlying models—to verify it demonstrates clinical knowledge, safety behavior, and generalizability for its intended use, with rigor scaled to the device's risk profile. Testing alone may not suffice; the FDA is also considering real-world evidence collection through standardized patient interactions, prospective clinical studies, or shadow deployment to confirm performance in actual clinical settings. The framework acknowledges that generative AI poses distinct risks compared to traditional software: devices can accept open-ended inputs, produce variable outputs, hallucinate plausibly false information, and evolve over time in ways that degrade performance unpredictably. The FDA has authorized over 1,000 AI-enabled devices to date, but most do not incorporate generative AI. The agency is accepting public comments through October 19 and has framed the discussion paper around targeted questions rather than binding guidance, signaling it is still determining whether competency-based testing, risk-proportionate postmarket monitoring, or other approaches best balance innovation speed with patient safety. The effort aligns with the Trump Administration's stated priority to accelerate AI medical product delivery to market.