Kaiju, an innovative asset manager and issuer of the AI-powered BTD Capital Fund (NYSE: DIP), is helping investors deploy AI in their portfolios and revolutionizing active management.
Pulse 2.0 interviewed Kaiju Worldwide CEO Ryan Pannell to learn more.
Ryan Pannell’s Background
According to Pannell:
“Kaiju started out as a private fund manager— Kaiju Capital Management— which used quantitative methods(including AI) to make investment management decisions. At this early stage, we managed several successful funds (all of which are now fully subscribed) using this technology for years, committing fully to AI in 2019.”
“Our team has been involved with AI and machine learning for over 30 years, so we’ve got a lot of experience with the technology and systems required to responsibly run an AI shop.”
Commenting on their venture into the ETF space, Pannell said, “About two years ago, we wondered why we weren’t seeing AI on the public fund side— specifically in ETF wrappers— given how powerful and profitable we’d found it to be on the private fund side. We found that there really weren’t any firms using fully autonomous AI. The reality is that AI systems are extremely expensive to set up, and require highly specialized personnel to maintain and improve, and the fees that ETF issuers can charge are generally too low to justify the heavy lift. But since we already had fully functional systems in place, it was a shorter path for us to set up a new RIA—Kaiju ETF Advisors— and launch our first fully AI-curated and directed ETF (ticker: DIP). And now, here we are; a global ecosystem of companies focused on creating, using, licensing, and selling AI-curated and directed investment strategies and products.”
But the path to AI-powered investment products was not linear. “My background is pretty diverse: I started in the arts (language and literature, film and photography), and spent time in veterinary medicine before making my way through web-based application development in the early 2000s. That led me to applied cryptography and some theoretical physics, and I spent the better part of a decade building secure electronic file delivery systems for global enterprises before turning those skills to quantitative equities trading. My performance working for a large hedge fund as an independent portfolio manager brought some attention, and that led to my first fund, which in turn led me to build a bigger fund, and along with it, Kaiju.”
Formation Of Kaiju
How did the idea for Kaiju come together? Pannell shared:
“Our company ideology came from collaboration with Kaiju with COO/CTO David Schooley to build an unconventional, science-based investment manager. We don’t see a lot of innovation in investment management. Pedigree is more important than performance or quality of thought, and that leads to stagnation and poor performance over time. We like to say that if you took a traditional fund manager and a Silicon Valley start-up and smashed them together, you’d get something like Kaiju. There are no beanbag chairs or pinball machines here, but we do work in a virtual, globally distributed environment across 17 countries and 13 time zones with an incredibly diverse workforce. We maintain an extremely high level of efficiency and don’t need to sacrifice speed for quality. Kaiju is a unique place to work, and so far, it’s been a very rewarding group of companies to build.”
Core Products
Commenting on Kaiju’s core products and features, Pannell shared,”The lifecycle here starts with AI-based IP generation. We take successful strategies and teach those to AI scientists. They then teach ‘the machine.’ We go back and forth refining outputs and creating guardrails that optimize profitability while keeping risk on a short leash, and when we’re ready, we determine whether the strategy is a better fit for public or private consumption. If the answer is “private”, then we run those strategies internally, in one of our funds managed by Kaiju Capital Management – or might choose to lease them to another fund or family office if there’s better shape to that. If the answer is “public”, then we make whatever modifications are necessary to launch the strategy in an ETF wrapper for public access, and that gets managed by Kaiju ETF Advisors.”
Evolution Of Kaiju’s Technology
Kaiju’s technology has evolved since launching. “Technology, as a whole, of course, evolves at an exponential rate, and AI is no different. From the quality of data we consume and train the models on, to the cost of the infrastructure, to the sophistication of our strategies thanks to leveraged new technologies, it’s an across-the-board improvement. The systems have become vastly more sophisticated since inception. The way in which we manage risk across market strata has changed as well; we run discrete engines at the underlying level, sector and industry levels, and broad market levels, and then all of those ‘talk’ to the Regime Classification and Regime Change Detection engines to ensure that the collective AI is making the most well-informed decisions it can. All that together forms the core of our ARC system (AI Risk Containment), and that’s a beautifully powerful system to watch work.”
Formation of the ‘Buy the Dip’ ETF
Kaiju’s flagship ETF, the DIP ETF came together over careful consideration”DIP wasn’t actually our first idea for an ETF. We originally wanted to bring our flagship strategy, BEX (Bilateral Equity Corridor Strategy), to the ETF community, but when we tried, it proved way too complex for market makers to work with. It’s completely market-neutral, risk defined, and has killer performance, but it requires complex options spreads and intraday rebalancing, and right now, that’s too much for market makers to handle. So we asked our partners what they would consider “easy” to work with, and they told us ‘long only, equity only’ to start. We went back to our toolbox, and we actually had a strategy that conformed to that (half a strategy really; the full version is dual-sided, so it contains the short side as well), and we modified it to fit in an ETF wrapper. Given that it predates on low to high mean reversions in artificially oversold stocks, we thought it’d be a great fit for investors who like the idea of ‘buying the dip’ but struggle with consistency. By bundling powerful AI they’re never going to get access to, with a relatively conservative strategy, we felt we were bringing the best of both worlds to an investor community who is just starting to consider trusting AI to make their investment decisions.”
Significant Milestones
Pannell pointed out some of Kaiju’s most significant milestones:
“The completion of our internal engines across the board the ARC certainly is a huge accomplishment. The ARC examines 220TB of data per trading day, and DIP alone makes 2 billion discrete examinations when considering the daily portfolio rebalance. It’s able to transact on the nanosecond scale,, and it’s completely self-evolving. While human portfolio managers take the weekend off, ARC and our other autonomous systems play millions of ‘games’ against themselves using 15 years of tick-level market data across the entirety of the US equity market and come back on Monday even smarter and more apt to market conditions. It used to take weeks to perform a retraining cycle, and now the shorter retraining cycles are performed overnight, and the longer ones are covered over the course of a weekend. That’s amazing to see. And obviously, working with Citadel Securities to bring the first fully AI-curated and directed ETF down to the NYSE floor is a big win for us.”
Customer Success Stories
After asking Pannell about customer success stories, he commented, “On the IP front, we’ve delivered returns to our foundational investors of over 900%, which is extremely gratifying; we take investor trust very seriously and are always looking to deliver outsized returns relative to risk. On the ETF front, we think that DIP is demonstrating that it is stable and has room to run. The fund launched six months ago and is currently executing completely on-model, as it spent most of the time leading up to the regional banking collapse outperforming the S&P 500. It suffered through that collapse due to a couple of key artificial disadvantages. Due to the wariness around a fully autonomous, AI-curated, and directed system managing the rebalance, we were required to not exceed a 20% per day rebalance, and our weighting was additionally limited at the sector level. These types of handicaps aren’t ideal for an AI system; when it “knows” it wants to dump the whole portfolio and reset, you really do need to let it. As a consequence, it took 15 days to do a rebalance it wanted to do in one. Those constraints have now been removed, and we’re looking forward to DIP’s autonomous learning system making the changes necessary to return to outperformance, as it was designed to do. That we were able to move past this to greater confidence on the part of our partners is definitely a solid success story for AI. On the private fund side, our latest iteration of BEX returned 6.4% on allocated capital last month (its first month of live operation since the rebuild of that strategy), so that’s a very rewarding performance to see as well.”
Funding
In terms of funding and revenue metrics, Pannell replied: “Globally, Kaiju has always been 100% founder financed; we’ve never asked anyone external for funding. That has been a successful model for almost half a decade now, though we may consider accepting a very small number of select new investors via a single round, as we’re looking to expand quickly with new products and more advanced iterations of our technologies. As for revenue, we continue to grow, advance, have no debt, and pay all our bills on time, so…I’d say we’re doing all right.”
Total Addressable Market
What total addressable market (TAM) size is Kaiju pursuing? Pannell assessed, “There are certain strategies we employ that have a max cap allocation attached. It wouldn’t matter how badly you wanted to add more money to them; they top out at a certain amount; there are finite opportunities that present themselves. Others, like the DIP ETF, don’t have a ceiling, so theoretically can take unlimited capital. Given that DIP seeks to outperform the S&P 500 using primarily the same components, ideally, we’d like to be considered as an add-on or alternative to one of the big S&P index ETFs— so maybe 5%— 10% of that weight. Globally, we’d like to see Kaiju hit $10B in valuation, and we think that’s doable given the strength and current value of our IP assets.”
Differentiation From The Competition
What differentiates Kaiju from its competition? Pannell affirmed:
“For starters, it’s the virtual, globally distributed workforce. We’ve been doing that since 2013, long before the pandemic made it ‘normal.’ The key advantage is that we are able to hire the best assets, period, not just the best assets in a particular region or who are willing to relocate to that region. The need to drag people into a physical office is ensuring that firms who insist on doing so are never ever going to be able to match our talent pool.”
“I’d say the next most obvious differentiator is the culture. Our culture is pretty attractive. It’s collaborative, challenging, dynamic, and supportive. We operate in a flexible and open office—we set and maintain a pace on a unique schedule because we hire people who thrive in this kind of environment. We are not a traditional fund manager, and we don’t run a traditional fund office.”
“Finally, another competitive advantage is our collaborative approach. The C-Suite here will listen to interns if they have a good idea, and there’s no hierarchy policy that prevents those ideas from being shared. Good ideas make it to the top quickly, and bad ideas get culled without the negative attributions which occur at so many other companies, and as a result, the idea churn here is optimized and always positive. Whether or not your idea made it, everyone here feels heard.”
AI in The Finance Sector
According to Pannell, the greatest roadblock to the widespread adoption of AI in finance is expressed in one word:
“Trust… We’re coming off the back of three years of snake oil in new technologies with the crypto boom/bust characterized by the FTX and Luna fiascoes, and investors are understandably jaded and gun-shy. Now, here we are with a new technology that (again) promises to do all kinds of powerful, valuable things, and the market is skeptical. All financial professionals need to gain investor trust before we can see increased comfort with what will be an ever-increasing application of AI in investment portfolios. There are two ways that we can achieve that: performance and transparency. First, of course, AI needs to prove it can do what we all claim it can: outperform humans in specific areas to add value. You can’t talk your way around continuous poor performance; you’re either delivering or you’re not. The second required component for regaining investor trust in bleeding-edge technologies is to treat the investor like a partner. Stop using deliberately complex ways of describing AI and the process that underpins it, and be honest about what AI can and can’t do— because it has significant weaknesses in some key areas. If you do those two things, and if you can prove that the value proposition is more than slick marketing and show that you acknowledge the challenges with global AI application, then you can rebuild the trust that vastly less responsible technologists have blown up over the last few years.”
Frustrations with AI in the Current Consumer Climate
What do you find most frustrating about the current consumer climate with respect to AI? Pannell answered: “The desire to see AI trip up, stumble, and fail. I mean, I get it; it’s natural for people to find joy in the demonstrated shortcomings of a technology touted as “the next big thing”, especially when the idea of that technology’s widespread adoption is disquieting to many. The idea of AI can be scary: Will it take over my decision-making autonomy? Will it render my job obsolete? Will it cause me harm in some way that I don’t expect? There’s a lot of apprehension, and so when AI proves itself to be fallible in some way, it pushes back the reality of any kind of universal adoption, and that brings some comfort.”
“What I wish we’d see instead is the acknowledgment that we’re still at the very early stages of what I’d call ‘useful AI’ and that this is a long road. There will be bumps along the way, but instead of celebrating AI’s shortcomings, it would be more proactive to acknowledge them, work to address them, and celebrate its successes instead. Because at the end of the day, applied responsibly, this is a technology that could be of huge benefit to all of us— but it will take time and patience.”
“So far, over the first half of DIPs life, it has outperformed the S&P 500, but I didn’t get a lot of calls or emails that said, “Hey, look at how great DIP is doing!” After the regional banking collapse, however, I did get several calls and emails along the lines of, “What’s going on with DIP? What’s it doing?” to which I replied, “Learning…” AI isn’t perfect in all scenarios, and that’s not a realistic goal at this point in time, but it can learn at a faster pace than humans can. You just need to give it time to do that. In the beginning, there are successes and failures; they make progress and regress. But there’s a tipping point at which the successes outpace the failures and become exponential rather than geometric in frequency. After that, the technology can take off like never before.
Future Plans
Does Kaiju have other AI-directed ETFs in the pipeline? Are they similar to DIP? Pannell disclosed: “We do have other AI-directed ETFs currently in development. Because it’s based on the S&P universe of stocks and it’s long-only, DIP was designed to be a “soft landing” for investors interested in trying out what AI-making endpoint decisions look like. We’ve had requests for ETFs that address specific pain points for both retail and institutional investors that we’re following up with: things like outsized alpha with a different risk profile to DIP and universal, holistic portfolio hedging, which is what we’ve been most approached about. Given that we have iterations of those solutions already on the private fund side, it just boils down to making the required changes to those solutions so that they fit in an ETF wrapper.”
Future Company Goals
“We want to continue building novel AI-related trading IP and, ideally, have a reliable mechanism for putting that to work for any investor who wants access. Currently, that access doesn’t exist. When it comes to investment management, AI is reserved for big funds and the wealthiest clients… But it doesn’t have to be that way. ETFs are a great way to leverage those technologies in a transparent, liquid way that anyone who wants to can take advantage of, but we are exploring others.”
