How One Firm Plans to Protect the Edge Investors Get Through AI

To prevent crowding, a Softbank-backed firm is limiting access to its latest AI signals.

Illustration by II

Illustration by II

When an investor gets an edge, whether through AI or a rare data set, it’s often a matter of time before its effectiveness is diminished.

One service provider in the business of prediction thinks it has at least part of an answer: cap the number of clients who can use it.

Qraft Technologies, a Seoul-based financial services company founded in 2016 and backed by SoftBank Group, has launched, an AI-powered tool that uses tick data — the minimum price movements in stocks and derivative markets — to find signals that will predict market-moving events. After analyzing terabytes of data on daily changes, the firm says that can identify even the tiniest irregular transactions that help its users spot things like mergers and acquisitions, earnings shocks, and FDA drug approvals prior to that information becoming instantly commoditized through press releases, newswires, and social media.

In addition to trades by company insiders and other investors, Qraft says it can also effectively detect “traces of smart money or informed traders” that subscribers can use to inform their investing. Stocks flagged by the platform have a 28 percent chance of signaling a corporate event, more than three times higher than random selection, Qraft said in a statement. The company did not disclose the cost of the service.

The service can level the playing field for “not only quant managers but all types of investors who need tools for data-driven decision-making for investment and risk management without necessarily spending a massive amount of resources that only a few would bear,” Francis Geeseok Oh, APAC chief executive at Qraft, said. Geeseok Oh is also the head of Qraft’s ETF business, which has three funds. The company, which raised $146 million from Softbank in 2022, said it expects to continue developing and launching new products supported by its AI research.


No person could digest even a fraction of the market data available each day. If they could, their analysis would be flawed, argues Marcus Kim, founder and CEO of Qraft. “Though markets remain uncertain and volatile, AI technology holds the power to deliver results beyond the limitations of human biases, accounting for mass amounts of data like that of the options trading market,” Kim said.

But Qraft has no intention of democratizing that power offered to an unlimited number of users. To “maintain the alpha” of, it will only be available to 1,000 hedge funds and other asset managers at the start, an uncommon approach, according to some observers.

“I am intrigued by this concept of limiting the use,” said Paul Fahey, head of investment data science at Northern Trust Asset Servicing.

Like other large financial services firms, Northern Trust has invested millions of dollars in collecting data, making it useful, and democratizing it, Fahey said. The company is exploring what data and insights it can share with smaller asset managers, helping to level a technology playing field and allowing them to focus on their investment acumen and skill. Limiting the number of firms it offers data services and insights to hasn’t been a consideration but Fahey can see why others might do it, he said.

Putting a cap on the number of customers using an effective tool would create exclusivity and allow a company to charge a premium for it, Fahey said. It might also preserve its effectiveness, at least until someone else is able to develop something similar.

“If everyone has access to the same data, then its ability to add differentiation or value would diminish, and would diminish over time, and that’s likely a fairly short timeline,” Fahey said. The same applies to any report or signal service. Scarcity makes them differentiating and there is always a possibility that someone else develops it.

Northern Trust, Fahey said, views data a little differently: the most valuable dataset an asset manager has is their own; the collection of information on markets alongside their portfolios, research and investment assessments. “I would also contend that that’s probably the most underutilized data set, in large part because it’s often not readily accessible because it sits in multiple places,” Fahey said.

“I’ve not heard of limits on subscription of data, but it may very well be out there,” said a person who oversees data platforms at a large bank and asset manager and was not authorized to speak publicly about its relationships with customers.

“I think the price for the data largely drives permit to play/adoption. The timeliness and the accuracy of that data is key and that drives the premium, this is especially true in the alternative/external data space.”