Is Delphia safe?
Delphia is a mobile-only, management fee-free, AI-powered automated investment advisory platform. Pitching itself as using an approach designed to emulate that of a modern quantitative hedge fund, Delphia says it plans to use its proprietary model to leverage users’ bank and Twitter data to inform the investment decisions that build its portfolios. The key phrase is “plans to.” Despite marketing that heavily implies otherwise, Delphia does not actually use the data it collects when users connect their bank accounts and social media profiles to inform investment decisions.
Also, 99 percent of Delphia’s assets under management (AUM) are attributable to the two large hedge funds it manages, meaning its interests may not always align with those of everyday investors.
Even if Delphia were to deliver on its promise of user data-driven investing, it still has a long way to go in proving that its novel approach can and will generate outsized returns. While on the surface appearing refreshingly unique, clean, well-intentioned and sensical in theory, Delphia is troublingly opaque, and investors can and should find better places to put their money.
Is Delphia legit?
First and foremost, Delphia is not a brokerage. Users cannot trade individual stocks or ETFs. Bonds,
options and other securities are off the table.
That said, Delphia is legitimate as an entity. Its parent company is registered with the SEC as an investment adviser, and its brokerage service providers, Alpaca and Apex Clearing, are members of the SIPC, meaning client funds are insured up to $500,000 per account — just as they would be in TD Ameritrade, Robinhood or any other American brokerage.
Nonetheless, the SEC requires a disclaimer cautioning that “registration of an investment adviser does not imply any specific level of skill or training and does not constitute an endorsement of the entity by the Securities Exchange Commission.” Basically, just because Delphia is a legitimate entity does not necessarily mean that its investment strategies beat the market or that its portfolios are suitable for all (or any) investors. It also does not necessarily mean that Delphia’s intentions and interests are what they make them out to be.
Especially when dealing with newer auto-investing firms like Delphia, investors should conduct ample research before committing funds to portfolios over which they have no direct control.
What makes Delphia unique?
As described on its Crunchbase, “Delphia is building the world's first investment adviser that allows people to invest their data alongside their money in order to improve their investment returns.”
Delphia’s differentiator is its collection of and (supposed) use of user data to inform its investment decisions. The app asks users to grant it access to their personal data, including but not limited to credit card transactions and social media posts.
An informational video on the company’s website makes the case for investors consenting to Delphia using their data:
“When you connect your data to Delphia, you join our community of investors, which lets our algorithm analyze everyone’s attributes and behaviors together in order to make predictions about you and everyone like you. So when enough people invest with Delphia, we can make predictions about the entire stock market.”
Bottom line: this is misleading, which we will get to later.
How does it work?
Delphia offers a set of four investment portfolio strategies from which users can choose to subscribe. The platform has zero management fees and a minimum investment of $10. The app claims to have “a Wall Street-caliber algorithm designed to help you achieve superior investment returns.”
The Flagship portfolio is an AI-driven actively traded U.S. equity portfolio that “uses consumer data to improve the predictions of companies’ performance.”
The other three portfolios — the low-risk Conservative portfolio, the medium-risk Balanced portfolio and the high-risk Growth portfolio — follow the company’s “Passive Plus” Portfolio strategy. Each of the three achieves a unique risk profile of the same underlying strategy, consisting of varying apportionments of fixed income and (not necessarily U.S.) equity securities.
It should be noted that clients can only be subscribed to one strategy at a time. While some robo-advisory platforms offer the option to customize their portfolio to be an apportionment of multiple strategies (e.g. 10 percent in Strategy A, 25 percent in Strategy B, 65 percent in Strategy C), Delphia users must choose to go “all-in,” ascribing their entire account balance to one of its four portfolios.
Delphia rebalances its portfolios by making trades in large batches, usually roughly one month apart. For users with small account balances, this results in dozens of fractional share trades at a time. Clients subscribed to the Flagship portfolio currently hold positions in 200 stocks. As a result, users can expect to receive several trade confirmations and/or shareholder voting prompts every week.
Strategies Overview
| Portfolio | Equities | Fixed Income | Inception | Theoretical Performance Since Inception | S&P 500 Performance Since Portfolio Inception |
|---|
| Flagship | 100% | 0% | August 2020 | +36.22% | +26.45% |
| Growth | 80% | 20% | June 2020 | +19.01% | +29.51% |
| Balanced | 50% | 50% | June 2020 | +10.30% | +29.51% |
| Conservative | 30% | 70% | June 2020 | +3.81% | +29.51% |
So what’s the problem?
To put it plainly, Delphia does not do what it leads you to believe it does.
Powered by a fresh $60 million Series A investment (from a group of investors that included the disgraced FTX Ventures), Delphia has craftily branded itself as the startup with the user data-driven investment strategy without actually delivering on that promise.
None of the platform’s strategies incorporate user data. I repeat: none of the platform’s strategies actually incorporate user data!!
It’s right there in the fine print: “Currently, although Delphia Tech and the Investment Adviser are collecting some data through Member Contributions, the Investment Adviser does not use Member Contributions to power its advice.”
Delphia would likely argue that the company’s platform does not explicitly say it is currently using user data to inform investment decisions. And while Delphia may not have made any materially false statements, some are grossly, and probably intentionally, misleading.
For example, the app’s Flagship portfolio description includes the sentence, “This portfolio uses consumer data to improve the predictions of companies’ performance.”
Imagine the scenario where a prospective investor visits delphia.com/data and is inspired when she reads, “Profit from your own data, for once.” She’s on board with the idea that her “data will power our algorithm.” Finally, she is told, “As a data-contributing investor, you will be helping us build a company to redistribute wealth and take control over the greatest asset in your life – your data.” She’s sold on the vision and decides to make an account. She completes the signup and deposits her money. She is prompted with a message asking her to contribute personal credit card transaction data, her bank details and her Twitter login. She connects all of those because she believe what she has read. When choosing a portfolio, she selects the Flagship portfolio because she likes how it says, “This portfolio uses consumer data to improve the predictions of companies’ performance.”
The average investor would not know that when Delphia says “consumer data,” it is actually referring to publicly available data — not the sensitive, personal data the client just contributed. They wouldn’t know that Delphia’s Form ADV Part 2A states that its suite of AI tools “corroborate and calibrate publicly available data … [to] inform its investment advisory services to some of its Model Portfolios.”
See the problem?
Any other problems?
Yes! Another company promising some vision of democratized investing, another company using that vision to help a few really rich people get richer.
According to Delphia’s latest Form ADV filing with the SEC, as of late October 2022, the company has over $192 million of regulatory assets under management. Upon closer inspection, of those assets, less than $2 million is attributable to 6,205 individual clients not classified as “high net worth individuals.” The other $190+ million is split between five pooled investment vehicles, two of which are Cayman Islands-based hedge funds with gross asset values upwards of $126 million and $63 million, respectively.
To summarize, more than 99 percent of Delphia’s AUM is held in two offshore hedge funds. This is the same company that doesn’t actually use your data to aid your investments as it implies, yet advertises, “Your data is valuable. Hedge funds buy it to give themselves an edge, but you don’t see a cent. Delphia wants to change that.”
Other Considerations
Delphia recently introduced a program designed to reward users who contribute their data. Users can earn PHI, an ERC-20 token, when they connect their bank, credit card or Twitter accounts. The current rate is 0.2 PHI per week. It is unclear what utility PHI will have, but Delphia promises they are “building a rewards program within Delphia that unlocks utility and opens up access to unique features within the ecosystem.” We will see.
There also appears to be a program in the works that “redistributes” fees paid by Delphia’s hedge fund investors to retail clients. It is unclear how this will work, as the program’s “Learn More” link is broken.
Delphia Review The Bottom Line
For investors looking to trade equities and other securities at their own discretion, Delphia is not the place to do it. For those looking for hands-off, management fee-free investing, Delphia’s collection of portfolios might seem attractive. But the fact that the platform deceptively does not yet deliver on its key differentiator — the promise that users’ data would be used to power their investment strategies — combined with the troubling conflict of interest stemming from Delphia’s hedge fund ties, retail investors looking for outsized returns can find better value elsewhere with more established robo-advisory firms.
Updated on 2/7/2023.
Leo Garkisch is an entrepreneur and freelancer based in Princeton, NJ. A quantitative value investor, he spends his time developing and deploying automated deep learning-driven algorithmic trading strategies that seek to capitalize on long-term disjunctions between equities’ market prices and their intrinsic values. In his free time, he enjoys playing and watching baseball and beating his friends in chess.
|