July 2026 could be one for the history books. It’s not because of the AI correction trade, seemingly coming out of nowhere. It’s not because a hedge fund that was reportedly up 400% year-to-date accelerated the sell-off and was bailed out by a much bigger hedge fund, reminiscent of LTCM, Bear Stearns, or Archegos. And it’s not because a broad market index like KOSPI with over 800 constituents fell -22% in a month just because 2 AI stocks accounting for nearly 60% of the index at its peak tanked in the midst of the correction, partly due to leveraged single stock ETFs recently launched on those names.
 
I believe the most significant event during July was the FX intervention coordinated by the Japanese Ministry of Finance and the US Treasury Department. The MOF had intervened several times this year. But it always reverted and the time it took for the yen to revert back to its previous levels became shorter each time, most recently intraday. Of course, there are plenty of reasons why the yen should be weak.
 
We started the year with the expectation of at least two Fed rate cuts during 2026. Recent fed funds futures imply one hike before the end of the year. Meanwhile, the implied rate for Japan by the end of the year is little changed from the start of the year (and this includes the actual +0.25% hike in June). While Japanese headline inflation has moderated this year, the PPI has sharply accelerated due to rising import prices from the energy disruptions and yen weakness. Wages continue to rise and consumers have now experienced 6 consecutive months of real wage growth. It is only a matter of time before businesses begin to pass on higher costs to consumers, driving CPI higher, especially as utility subsidies are expected to expire this September and fuel price caps begin to taper (unless the government can find another source of contingency reserves). The utility subsidy alone is said to increase core CPI by +0.4%~+0.6%. And yet, the BOJ remains on the sidelines. The market, justifiably, believes that they are behind the curve.
 
Additionally, the Takaichi administration’s fiscal expansion and growth policies have begun to pressure long-term rates. While she insists that she will not finance with excessive debt, she has been equally unclear where the financing will come from (while simultaneously discouraging the BOJ from tightening sufficiently). Add to this the temporary consumption tax cut on food from 8% to 1% for the next 2 years and one can understand why Japan’s fiscal risk premium is rising.
 
Within that backdrop, we saw the first coordinated intervention since the Great East Japan Earthquake in 2011. However, at that time, it was to stem the yen surge. The last time the two countries worked together to buy yen was during the Asian Financial Crisis in 1998. Not only was it publicly acknowledged, Treasury Secretary Bessent went so far as to highlight the Fed’s FIMA repo facility as a mechanism that Japan can use for future intervention, later confirmed by Finance Minister Katayama (presumably to highlight that Japan can continue to intervene without selling its US treasury holdings).
 
This intervention directly attacked one of the most crowded trades in the world, arguably more than the AI trade, utilized not only by hedge funds and financial institutions, but even Japanese corporates and retail.
 
Now, let me say for the record that I have never gotten FX calls right. I suggested my wife-to-be to short the yen at 82 but her levered trade got stopped out while we were on our honeymoon during the earthquake, driving the yen to 78; we were scuba diving in Maldives during the earthquake so we heard about it much later. I’ve been expecting the yen to strengthen for the last 2 years until Takaichi-san came to power but still thought it wouldn’t weaken much further … and, of course, it did.
 
But maybe, just maybe, the carry trade will reverse. At the minimum, with 10-year JGBs above 2.8% (and the 30-year near 4%), one would presume that at least Japanese life insurers and pension plans might raise their domestic fixed income allocations. Finance Minister Katayama has said that she wanted to pursue measures to encourage pension funds, including the GPIF, the largest sovereign wealth fund in the world, to increase investments in Japanese assets. She has also discussed including JGBs in the NISA program, Japan’s tax-exempt retail investment program, as well as possible inheritance-tax changes to make JGBs even more attractive for households. And finally, my conspiracy theory is that Japan’s obsession with stablecoin reforms is attempting to create another stable buyer of JGBs. Through repatriation of the yen, I believe the Takaichi administration is attempting to create willing buyers of JGBs in order to offload the massive JGB holdings by the BOJ (reportedly owning 50% of total issuances).
 
Of course, this doesn’t resolve Japan’s debt servicing problem. But it does buy them time for Takaichi’s growth policies to kick in and nominal GDP to accelerate, thereby increasing tax revenues and to start servicing that debt.
 
Despite all of that commentary, it actually doesn’t matter to us one way or another except to reinforce my views and I feel highly confident about the growth potential of Japan. Most outside investors probably think of Japan as a market for long/short or activist strategies. In fact, I recently met a first-time prospect and, as I explained that we aren’t an activist, they kept asking “how do you get a cheap stock to fair if you aren’t an activist?” I was tempted to ask, “are all of your global managers activist?”. I guess they’ve never heard of stock-picking or active management.
 
I suppose Japan is still not viewed as a growth market. Takaichi-san must work harder to get the world to notice that there is much more to Japan than just “better governance”.
 
Masaki Gotoh
 
PS: Since beginning the draft of this monthly, the yen has since reverted about halfway between pre- and post-intervention … of course.
 
*** We will be co-hosting the first inaugural SOHN Tokyo Investment Conference on Oct. 16. We hope you can all attend and help support a great cause with the Karen Leung Foundation *** ​​​

​

===

​

​​“There are no experimental failures. There’s only more data.’” – Bryce Lynch, played by Chris Young, from the Max Headroom television series (Season 2, Episode 2 “Deities”).
 
Television programs in the 80s were great. We had hilarious sitcoms like Three’s Company, Happy Days, Night Court, and Cheers (well before censors made sitcoms less funny, at least from a GenX perspective). It was action-packed with programs like the A-Team, MacGyver, and The Equalizer. I could go on and on, but these are just the ones I own on DVD.
 
We also had great science fiction programs well before computer graphics was mature (just as a side note, there were no GPUs and no NVIDIA back then, but we had “video cards” from IBM. I still remember when I was able to upgrade mine to VGA, which could show a whole 256 colors simultaneously! I thought it was the coolest thing).
 
And the technology they prophesied were well ahead of its time. We had flying vehicles from “V”. Granted, it was by reptilian aliens trying to eat us. That might sound stupid, but the show taught us about the ease with which an ordinary, democratic society could become fascist, which might have some parallels today. We had not just autonomous cars but talking ones from “Knight Rider”. K.I.T.T (Knight Industries Two Thousand) could drive for you and talk to you while you played your favorite Atari video game. We had, of course, many exciting traditional science fiction programs like Battlestar Galactica (I own both the original DVD and the remake which I hadn’t seen in real-time. This is a rare case when I thought the remake was better).
 
But my favorite Sci-Fi was “Max Headroom” which featured artificial intelligence via a computer-generated character created from the mind and memory of a star reporter who, during one of his investigative reports, crashes his motorcycle against a signage which read “Max(imum) head room 2.3 meters”. While in a coma, the bad guys needed to know what he knew so they digitized his brain and accidentally created the self-aware, artificial construct called Max Headroom. It is set in some unknown, Blade Runner-esque future in a dystopian society ruled absolutely by a group of television networks, with the most powerful one (defined by consistently high ratings) being Network 23 where Max Headroom would later become a popular host.
 
Bryce Lynch is the genius hacker who heads R&D at Network 23 and the one who inadvertently created Max. Being only 16 and in front of a computer all day, he doesn’t know right or wrong (at least at the start of the season). Thus, the quote above (although the context within which he says it isn’t as profound as the quote itself, especially after the recent AI-driven hacks).
 
This Sci-Fi program, which only ran for 2 seasons, had correctly identified several of the social issues related to technology that we face today. There are several episodes that delve into digital addiction from media using algorithms to stimulate neurochemical dopamine loops in order to maintain user retention and commercial monetization, similar to what we are seeing with social media today. There are some that focus on media that benefit from outrage and conflict coupled with video manipulation to manufacture consensus, just as, in the present, engagement-driven algorithms promote polarizing content using controversy to increase views. Furthermore, there has been greater concern about generative AI deepfakes and disinformation, just as this 80s program had shown. There are several episodes that I particularly liked which showed how people were being diminished to data points leading to legal verdicts, credit controls, or access restrictions of societal services, all automated through AI. In the opening sequence, Bryce Lynch says “you’re looking at the future … people translated as data.” And almost all shows are based on continuous two-way monitoring as if that were normal where user behavior, preferences, and psychological responses are constantly tracked, just like smart devices, data logs, browsing habits, and biometric signals are being used to create behavioral profiles, converting personal privacy into a traded commodity.
 
Like Bryce Lynch (but unlike Max Headroom who is comically very “human”), AI is amoral (not to be confused with “immoral” which is a human trait). And that makes it dangerous, much more so than humans. Instrumental convergence says that, regardless of the goal that a superintelligent being pursues or is asked to pursue, there are similar sub-goals implicit in all goals, the obvious one being self-preservation; one can’t complete one’s goal if one is dead (or turned off). Does that mean all goals will ultimately lead to an existential risk for humans? Or what about the Midas problem based on the Greek myth of King Midas who wished for everything he touched turn to gold, leading to his starvation. Put simply, humans cannot explicitly write all of the rules that account for every possible case. And therefore, as it gains superintelligence, is total control of AI possible? Or could we at least align it to our shared values and ethical standards? But what does that even mean? Aside from “Thou Shalt Not Kill”, I can’t think of any shared value that we could agree is universal. And even with this commandment, what about physician-assisted suicides or euthanasia? I doubt we can all agree for or against mercy killings.
 
AI ethics is an extremely complex topic. I hope they find a solution before Singularity.
 
In the meantime, I’d be happy if they could just get Copilot to stop hallucinating …​