Exchange Traded Concepts (ETC) was the first company to offer white-label services to aspiring issuers and the company is still very much a presence in the industry -- even though it's not as prolific an issuer as its white-label counterparts. ETC recently worked with existing fund sponsor partner Qraft to launch Qraft's fourth fund on the platform, the Qraft AI-Enhanced U.S. Large Cap ETF (QRFT) .
You'll notice this is yet another fund that looks to leverage the power of so-called artificial intelligence. For what it's worth, similar to the fund and firm I reviewed recently, Qraft has some serious AI chops, recognized by not just me, but by the folks at Softbank who made a $146 million investment in the company at the beginning of 2022. Let's take a look at Qraft and the new fund.
Quantitative Craft?
Per South Korea-based Qraft's website, "The name Qraft is an amalgamation of the words 'Quant' and 'Craft,' conveying our purpose in crafting quantitative solutions enabled by our proprietary AI methods." The company was founded by a group of quantitative traders who found that over time their trades were becoming increasingly crowded as the rest of the world began to match their capabilities.
As with anything data-related, quality results come from strong analytics but start with a clean dataset and a clear way to access that data. Qraft sources company financial and macroeconomic data from a number of providers as well as what is known as "unstructured data" like press releases, earnings call transcripts, and the like. The AI-powered funds I've reviewed in the past all have some secret sauce algorithm (that may or may not involve ChatGPT?). Qraft took the step to create and patent an entire Application Programming Interface (API), the Kirin API platform, which allows them to take in data from providers like Compustat, Thomson Reuters, Federal Reserve Economic Data (FRED), and Nasdaq's Quandl platform and proceed to scrub, align and standardize into an optimized dataset. The result is a data source that drives security selection in all of Qraft's ETFs as well as the company's other business lines which include security selection, asset allocation, and electronic trading services as well as an AI-driven risk indicator.
QRFT
This is the fourth fund from Qraft brings the total assets under management to just over $20 million. QRFT is an actively managed fund, so the prospectus is the place to read about the security selection process. Because of the "black box" nature of the process, there isn't a lot of insight that can be gleaned here although the website does have a page outlining the steps they take in that process. Still, we can learn that the fund has a target number of holdings between 300 to 350, each of those holdings must have a market capitalization of at least $4 billion and the portfolio is reviewed and adjusted at the beginning of each month. The portfolio allocation methodology is simply described as "weighted pursuant to a methodology designed to maximize risk-adjusted return."
The language in the prospectus doesn't differ too much from a traditional non-AI actively managed fund process description. It talks about managing factor, volatility, and other risks. The only big difference is it points to things like "Bayesian neural networks" instead of a "portfolio manager's X years of experience" as the way they will achieve their strategy's goals.
Wrap It Up
Qraft's other funds include a momentum product, a value product, and a balanced strategy, all of which have a clear directive to either maximize or minimize a relatively narrow range of variables. This new fund has a much broader mandate, and I am interested to see how the process ends up performing. To be clear, Qraft's approach is not "set it and forget it." Throughout the site and literature, the company talks about how AI implementations do all the heavy lifting, but the final step is human-led review and implementation. The company has had success in its other lines of business selling into the institutional investing space and QRFT will be going on my watchlist -- will it also succeed in the retail ETF space?