Despite the relative strength of the broader index, particularly the mid and small caps that were up +3.8% and +5.1% respectively, November saw a huge reversal of the N/T ratio with the tech-/exporter-heavy Nikkei 225 down -4.1% and the broader Topix index up only +1.4% with the large caps flat for the month. While this bifurcation (and reversal of the N/T ratio) was significant, the ratio is still at very elevated levels compared to history and had only reverted to that just less than 2 months ago. It feels like the Nikkei 225 is being driven more by AI/tech than the US indices. In fact, the Nikkei 225 (in local currency) has outperformed the S&P 500 and Nasdaq year-to-date and is trading inline the SOX index since the second half of the year. Of course, some of this is due to the Takaichi trade since the spreads between the US indices and the Nikkei 225 were flat up to September end (as was Topix which is less tech-sensitive).
 
However, in actuality, the Nikkei 225 doesn’t have all that many tech names, much less AI plays. About 1/3 of the index by weight is classified as electronics or precision but these arbitrary industries include names like Terumo and Olympus (both medical) or Fanuc and Keyence (machinery) and a slew of consumer and industrial electronics conglomerates that are anything but high-tech. Conversely, there are names like ShinEtsu Chemical who derives half of their earnings from wafers but is classified as a Chemical company or Fujikura which has become synonymous with data center investment which falls under non-ferrous metals. And the closest true AI name, Softbank, is still considered a Telco company despite having spun out Softbank Mobile (both of which are included in the Nikkei 225). Net-net, I’d guess that the total exposure to high-tech from an earnings perspective is probably pretty close to 1/3, with 70% of that third coming from Advantest, Softbank, and TEL.
 
So, the fact that the Nikkei was down -4.1% when the SOX was down only -2.8% feels like a lot (but then again, in October, SOX was up +13.5% but the Nikkei was up even more at +16.6%). And this is despite the fact that the yen had weakened even further due to our highly dovish PM and the recently announced supplementary budget which includes a massive stimulus package, sending long-term yields to an 18-year high. Still, in absolute terms, the 10-year is at 1.9% with only Switzerland having a lower yield among major developed countries (I’m not sure if it’s still fair to call China an emerging market, but their 10-year is slowly lower at 1.8%).
 
While most of our larger holdings were firm, the bulk of the underperformance came from what was our #3 SPE company that I mentioned last month. It had organically grown from a below average position size to our 3rd largest since we began our investment on Liberation Day. It ran as high as +71% from our cost but it fell -15% last month. I mentioned that I’d add on dips and it almost got to my entry level near the end of the month but had decided to stall from there. To be honest, I’m not sure if I can get it at the level I’d like, but we still have other names that we are still building which feel more attractive from a valuation perspective at present.
 
As always, we’ll keep an eye on relative attractiveness but, despite the sharp underperformance, I’m fairly comfortable with our positioning and look to continue to rebalance on dips as I’m sure the volatility will continue.
 
Masaki​​​

===

​​

“The company is not currently engaged in any substantial business activity and has no plans to engage in any such activity in the foreseeable future.” – quote from a Bloomberg article (Feb 7, 2000) regarding the SEC filing of NetJ.com.

 

I remember reading about NetJ.com in 2000 … in retrospect, I guess it was the precursor to backdoor listings and SPACs, although nothing as structured as it is today. This was a listed company who described themselves as a business that does nothing and will continue to do nothing. And yet, it was trading at around $20 mln at the time according to my old notes (and was already a 5-bagger when I read about it). Looking it up on Bloomberg now, it appears to have successfully completed what a back-door listing vehicle should do and had made acquisitions to turn itself into Zoolink Corp (a developer of next generation “Intelligent Internet Data Centers”), then Action Energy Corp, then SMC Recordings, then SMC Entertainment, and most recently in July of this year to Fyntechnical Innovations with a market cap of $176k and apparently “provides an AI-based research platform for researching and trading different asset classes” (from Bloomberg). I don’t know what the companies in between did, but it appears that SMC Entertainment (and presumably SMC Recordings before it) was classified as a Music Record Label before later becoming an “incubator company focused on acquisition and support of … Fintech companies”.
 
I’m sure many of you remember this period (or at least anyone in their 40s or above). But it was a time when listing a company with a “dot-com” or even just adding “dot-com” or a “net” to your name was not only fashionable but profitable as valuations often rose, at least for a period of time. NetJ.com did both. But by 2002, the reverse trend began where companies who dropped the “dot-com” from their name would see their depressed stock prices rise. I’m sure many remember the demise of firms like pets.com. A similar phenomenon occurred with cryptocurrency by changing your corporate name to something with “crypto”, “coin”, or “chain” in it (as far as I know, the reverse has yet to happen).
 
But I have to admit, I don’t think I’ve seen a similar trend in Japan. In fact, going back to 1997, I only found one company that did so, a firm called Bunkahoso Brains that changed  its name to Brain.com in Dec 2000. Just 8 months later, though, they changed it again to Digit Brain (and later to Megabrain before being acquired by SBI Holdings). Maybe it was seen as tacky to use the “dot-com” moniker in your corporate name in Japan. But there were several companies who added “net” in 2000 such as JC Foods to JCFoodsNet (who later dropped the “Net” in 2003), Sun S to Vital-Net (who kept the name until delisting), and AM Japan to Digi Net. However, unlike the dot-coms in the US, none of these companies did anything remotely related to the Internet (or at least not according to the most recent descriptions). JCFoods/JCFoodsNet (who is now known as Delsole) produces frozen pizza, Sun S/Vital-Net distributed medical products at the time of delisting, and AM Japan/Digi Net was a wholesaler of printers that also sold used cars and was acquired by an automotive retailer a few months after changing their name. I suppose a “network” can have a broader meaning and perhaps was seen as less crude; we all know how Japan likes to try to be subtle.
 
As part of our research process, our investment team has a quarterly “Shikiho check” session, one of my favorite regular meetings. The “Kaisho Shikiho” is a 7”x5” quarterly publication over 2,000 pages long which is essentially a comprehensive database of every listed company in Japan, describing 2 companies per page. It contains basic information such as company description, technical and fundamental info, valuations, management and shareholder info, to name just a few (there’s actually quite a lot more in the tiny 3.5”x2.5” space per company; it’s so small that I, with my worsening eyesight, have to buy the “wide” version which is about 1.5x bigger and weighs about 1kg). The descriptions include a brief note on what’s been going on since the last quarter and provides independent PnL estimates for the next 2 years (albeit pretty rough estimates, but who can blame them since they update these estimates every quarter for nearly 4000 companies) . It’s an extremely efficient way to get a very quick, holistic view of any listed company in Japan. We go through portions of this book every quarter and review old and new companies, generally based on a pre-screened theme or two. We try to keep an open mind on both the theme and the companies themselves, even if we’ve kicked them out in the past. Maybe, after years of complacency, the company is restructuring some business lines, or they found religion and started to think about balance sheet efficiency. Maybe there’s a new shareholder or a change in management that piques our interest. Or maybe it’s simply that the stock had been falling and is now at interesting valuations. Whatever it may be, we discuss any that caught our eye and potentially place them on a list of names that we may want to start some basic desk research.
 
What I’ve noticed is the new name change fad has returned using “AI” and, to a lesser extent, “Robotics”. As far as I can tell, the first one in (listed) Japan was a firm called Ubiquitous Corp that changed to Ubiquitous AI Corp back in 2018. It’s a $25 mln company that’s been around since 2001, and they appear to be some sort of software company for manufacturing firms. In 2019, Shinkawa, a leading wire-bonder company, was acquired by Yamaha Motors and changed its name to Yamaha Robotics Holdings. I suppose one could call say that a wire-bonder is a robot company (as would be true of any SPE in that case) and, to be fair, Yamaha Motors does make industrial robots in addition to being the second largest manufacturer of motorcycles in the world. There have been other similar name changes like Robot Home which appears to be a real estate management platform company and GMO Research & AI (part of the GMO Internet Group) which is an online marketing research firm.
 
Similar to the dot-com era, many recent IPOs also have AI in their name. There’s an AI Cross, AI inside (I don’t know what they do but they have a cool domain “inside.ai”), AI Mechatec, monoAI technology, Aidemy, Laboro.AI, and one that is simple and to the point, AI Inc. But unlike the “net” name change, most of these appear to actually be AI-related businesses. They all utilize AI in some fashion (though one could say that about any company, just like every firm had a dot.com domain since the 2000s bubble). My favorite one has to be “Ai Robotics”, a $475 mln company that, according to Bloomberg, “provides direct to consumer planning and sales using artificial intelligence, artificial intelligence marketing services, and other services”. The description itself sounds awfully repetitive. A quick glance at their website tells me that they provide direct-to-consumer marketing tools, primarily to cosmetics firms (I’m not sure where the “Robotics” comes into play). While they initially had successful installations at some major Japanese skin care brands, they changed their business model to start their own skin care product and a beauty consumer electronics brand utilizing the software that they developed. And so, despite its name and the Bloomberg description (and its home page), it’s a skin care brand (and categorized as a “chemical” company). To their credit, they’ve been successful building the two brands in a highly competitive industry. I used to cover the Japanese cosmetics industry, so I know the segment fairly well. As such, I suspect their software works pretty well, and what better way to prove it than to start a brand on your own. I would think that selling the software is more sustainable, but I can understand the lure of growing a skin care brand, especially if it’s under your own eCommerce site like they do. If you can get scale and efficiently reduce the costliest element of cosmetics, being marketing, it is a highly profitable business.
 
But what I found most interesting is that, already, the reverse trend is happening. GMO Research & AI, after changing their name from just GMO Research in May 2024, dropped the “AI” last month to become GMO Product Platform, as well as changing their recently updated domain from “gmo-research.ai” to “product.gmo”. To be precise, they merged another GMO group company called GMO Town Wifi which is one of the leading free wifi service providers in Japan. I’m not entirely certain what synergies these two businesses have, but what I find particularly interesting is that marketing research is probably a prime example of where AI adoption will likely accelerate (similar to, presumably, what Ai Robotics had done on their own). GMO Research should be flaunting the (potential) AI capabilities, not hiding it. And the GMO Internet Group is, unusually for Japan, a company that is not known to be “subtle”.
 
Of course, it’s probably a company specific event and not start of a trend. While we could argue about valuations, I am a firm believer that AI and robotics (which is just a form of physical AI) will change the world, just as the Internet did since the dot-com boom. The valuations at its peak were just a decade ahead of its time. But, even if just a coincidence, the name change is timely, given the recent worries about AI valuations, ROIs, and circular AI investments. Neither GMO Product Platform name change nor the recent AI-related stock volatility questions my faith, of course. But I always like to keep an open mind about everything, and I will keep an eye on both trends.
 
In the meantime, someone should start a new firm and name it Crypto AI Robots & Co. It’ll probably be worth at least $20 mln.
​