Andrew here. Here’s a dilemma for the C-suite: Should employees be allowed to place bets with prediction markets?
I’m not talking about insider trading of your own stock. Imagine logistics managers using their unique supply chain visibility to wager on inflation data. They aren’t betting on their employers — but they are monetizing an information edge that the public doesn’t have.
Wall Street is already clamping down. Goldman Sachs and others are barring employees from trading financial and political event contracts, citing compliance and conflict risks. Let me know what you think the rules should be.
Goldman’s limits on prediction market betting
Prediction markets like Kalshi and Polymarket have become a big business, with aims of becoming a next-generation Wall Street.
But existing financial giants are limiting what their employees can do on these betting platforms, underscoring growing fears of wagers based on illicit information, Michael de la Merced writes.
Goldman Sachs set strict limits on employee bets on prediction markets, DealBook has confirmed. In a recent memo to employees, first reported by Bloomberg, the bank barred workers from wagering on essentially anything except sports and entertainment.
It’s tied to Goldman’s policies ensuring that workers don’t trade on material nonpublic information and that they avoid real or perceived conflicts of interest.
Other Wall Street banks are also addressing prediction market use. JPMorgan Chase recently sent a memo to employees that was less prescriptive than Goldman’s, urging workers not to “engage in transactions that create the appearance of conflict, or use information you learn through work for personal gain.”
A Morgan Stanley representative told CNBC that the firm had policies about prediction markets but didn’t elaborate. And Bank of America is reportedly in the process of publishing policy updates.
It isn’t just banks that are taking stands. The big hedge funds Point72 and Balyasny have barred employees from trading on prediction markets via their personal accounts, Bloomberg previously reported.
Insider trading concerns have continued to dog prediction markets, including after suspicious winning bets on geopolitical events like the U.S. military operation in Venezuela. Regulators have brought charges against a Google employee and against a U.S. Army soldier who made bets on Polymarket.
Kalshi has noted that it is regulated in the U.S. and promoted its efforts to crack down on improper trading. Polymarket, whose offshore market has been a focus of scrutiny amid claims of suspicious trading, said it was “committed to maintaining accurate, fair and transparent markets.”
But Wall Street remains keenly interested in prediction markets, even as financial titans are trying to figure out the regulatory outlook for the platforms. Firms like Susquehanna International Group, Jump Trading and FalconX are helping institutional clients use prediction markets for hedging and other trades.
And Kalshi has begun thinking about an eventual I.P.O. The Information previously reported that Kalshi had encouraged banks that might advise on the deal to work with its platform to bolster its institutional trading business.
HERE’S WHAT’S HAPPENING
Oil dips on new demand fears. The International Energy Agency said on Friday that oil demand would drop significantly this year. The high prices and supply disruptions caused by war in the Middle East are part of the problem, as is a slowing global economy.
A stricken New York City skyscraper puts the office-to-apartment conversion boom in focus. Developers around the country are facing questions about such projects after structural damage at Pfizer’s former headquarters in Midtown Manhattan has put the skyscraper at risk of partial collapse. The projects could now face a “Pfizer premium” involving more expensive safety reviews, insurance and funding.
Investors are closely watching SK Hynix’s U.S. trading debut on Friday. The Korean tech giant raised $26.5 billion, a near record, in a share sale on Thursday. It’s another sign of robust investor demand for companies at the heart of the memory chip boom driving the artificial intelligence trade, but broader bubble concerns haven’t faded.
The A.I. A-Team
Artificial intelligence is going to turbocharge the economy, wreck the job market or fall short of the hype.
Two prominent new appointments in Washington and Silicon Valley signal the stakes of correctly determining the technology’s effects on the economy and beyond. They involve Marc Andreessen, a billionaire venture capitalist and backer of President Trump, and Ben Bernanke, the former Fed chair.
The Fed named Andreessen to a “productivity and jobs” task force. The group will assess the economic impact of new technologies such as A.I. to inform central bank policy.
-
Andreessen will be joined by Charles Jones, an economics professor at Stanford who is currently on leave working at Anthropic, and Asha Sharma, the C.E.O. of Microsoft’s Xbox unit, which just announced a reset involving big layoffs.
The productivity and jobs task force is one of five created by Kevin Warsh, the new Fed chairman.
Warsh is bullish on A.I.’s potential. He believes it will drive up productivity, spur growth and eventually lower inflation — potentially setting the stage for lower interest rates. Fed officials mostly agree that A.I. will bring economic benefits, but they also fear that rapid corporate adoption of the technology could disrupt the labor market and that the debt-fueled A.I. infrastructure buildout could raise inflation in the near term.
Could a task force involving Andreessen, a noted “techno-optimist,” change policymakers’ views on A.I.’s benefits?
And then there’s Bernanke, who is joining an independent advisory board for Anthropic. As the Fed chief from 2006 to 2014, Bernanke helped steer the economy through the global financial crisis that defined his tenure.
Dario Amodei, Anthropic’s C.E.O., cited that experience on Thursday. Bernanke’s judgment “will make us better at anticipating and responding to how advanced A.I. affects work forces and economies around the world,” Amodei wrote in a statement published to the company’s blog.
-
In other A.I. news, Fidji Simo, the No. 2 executive at OpenAI, said on Thursday that she was leaving the company for health reasons, another complication for the A.I. lab as it prepares to go public.
Trump’s trades
President Trump’s recent financial disclosures pointing to giant earnings while in office have stunned his critics and Wall Street alike.
Now, The Wall Street Journal has revealed a critical piece of Trump’s moneymaking strategy: frenetic stock trading, much of it in his brokerage account at Charles Schwab.
A legal victory helped. Much of the trading began shortly after a New York appeals court threw out a roughly $500 million penalty that the state’s attorney general, Letitia James, had won in a fraud case against Trump family businesses, The Journal reported, citing unnamed sources.
Trump, who disclosed about 21,000 trades last year, used a direct-index trading strategy. That often involves buying and selling individual stocks in an effort to mimic the swings of a broader stock index. Some investors prefer this way of trading, in part because it can help offset a capital-gains tax hit.
The president’s family has investment accounts with other firms, including UBS and JPMorgan Chase, The Journal reported, citing anonymous sources. But Schwab stands out.
From The Journal:
Schwab is a dominant competitor in the brokerage industry and a pioneer in low-cost trading. Unlike banking giants such as JPMorgan and Capital One, the firm didn’t face accusations from the Trump family that it had “debanked” them, or closed their accounts, in response to the Jan. 6, 2021, riot at the U.S. Capitol.
Schwab founder Charles “Chuck” Schwab, who has occasionally appeared at the president’s side, has advocated for government-backed retirement plans, including a mandated savings program requiring employers to contribute to workers’ accounts, people familiar with the matter said.
Chart of the day
Venture capital firms poured $412.7 billion into start-ups in the first half of the year, more than any full year in the past decade, according to a new report by PitchBook and the National Venture Capital Association.
But aspects of that boom are stirring up concern in Silicon Valley and beyond:
-
Artificial intelligence deals — including a $122 billion mega-round for OpenAI and a $65 billion one for Anthropic — drove much of the frenzy, accounting for 86 percent of the investments. But that hyper-focus concentrates risk, the PitchBook authors write: “Even if A.I. delivers, returns will concentrate in a small number of winners, so funding a wide field at elevated prices leaves many of those entries exposed.”
-
Nearly 90 percent of the V.C. money invested in start-ups in the first six months of 2026 went to deals valued at $100 million or more, another kind of concentration risk.
-
The future of the V.C. fervor depends largely on how the forthcoming I.P.O.s for OpenAI and Anthropic perform. “A lot is riding on those two companies,” Kyle Stanford, a co-author of the PitchBook report, told DealBook.
Talking A.I. with the managing partner of Bain Capital
Every week, we’re asking a leader how he or she uses artificial intelligence. This week, David Gross, who leads Bain Capital, told DealBook’s Sarah Kessler that it’s difficult for companies to transition from personal productivity to change at scale. The interview has been condensed and edited.
How do you use A.I. personally?
It’s the most confidential adviser I could ever hope to have.
What have you told your direct reports or employees about how you want them to use A.I.?
It’s really been about not approaching it from a technology way of thinking, but as an imagination problem. You don’t really get that far by incrementally making something a little bit more efficient or easier to do.
Usually that sends you on the path of a big I.T. project, versus doing something in a more practical way.
Bain Capital invested in the Deployment Company, a joint venture with OpenAI and other investors that sends consultants to help businesses use A.I. software. What have you learned?
I think businesses are all experiencing the same challenges in terms of implementing this beyond personal productivity and implementing it at scale.
Do you see these forward-deployed engineers as a permanent need or just a way to get over an adoption hump?
It does feel like there’s a bottleneck that needs to be cleared because the technology is new.
When you are taking this very broad and powerful technology and implementing it in real-life companies, changing how business processes really work, I don’t see a technology solution for that.
THE SPEED READ
Deals
Politics, policy and regulation
Best of the rest
We’d like your feedback! Please email thoughts and suggestions to dealbook@nytimes.com.
