So much has happened in March already, it’s hard to isolate last month but I’ll try to focus the commentary on what happened in February only. Geopolitically, it was (comparatively) quiet but in Japan, we had our record-breaking decisive LDP win driven by the huge popularity of Sanae Takaichi; she was formally reappointed as Prime Minister on Feb 18. And the equity markets welcomed her reelection with a massive +10.4% return by both the Nikkei 225 and Topix. Among factors, there was a definitive value, small-cap tilt which is rather unusual during such a significant beta event. Japan was, again, the best performer among the G7 countries … in fact, for Feb 2026, Japan was the best performer among all developed markets (as measured by MSCI) and the 4th best among the all-country indices (after Korea, Thailand, and Taiwan).

 
However, looking under the hood, it wasn’t quite as broad-based as such a large move would normally suggest. The fact that value and small-caps outperformed is already an anomaly. But sector dispersion was incredibly high. In fact, using a crude measure of dispersion by computing the standard deviation of returns across the 33 TSE sector indices, monthly return dispersion was the 12th highest level in the last 35 years since the 1980s real estate bubble burst.
 

 

 

​

​

​

​

​

 

 

 

 

 

 

 

 

(Source: Bloomberg computed by TriVista Capital)

 

Among the top 15 dispersion months, 7 were around the dot-com bubble crash and 4 were around GFC. One was during the first month of COVID-19 and another was after the Asian Financial Crisis which drove the first city bank, Hokkaido Takushoku Bank, and Yamaichi Securities, one of the “Big Four” securities firms, to collapse in Nov 1997. The 7th month (Aug 2021) was due to a statistically anomaly where one of the 33 sectors, Maritime Transport, rose 50% in one month due to rising container shipping rates, ultimately to over 7x pre-COVID levels making new all-time highs, amid the sudden demand rebound as the world reopened and unprecedented port congestions and vessel delays. Removing this one sector (which is essentially made up of just 3 major Japanese shippers), Aug 2021 dispersion was around average levels during the 35 years. To be fair, this month also had one sector, Non-ferrous Metals, that stood out and was up +48% but even if we were to exclude the best and worst sector each month for the last 35 years, February 2026 made the top 40 dispersion months (of which 20 months were around the dot-com crash, 7 were around GFC, 4 after Abe-san’s landslide victory (and the introduction of QE), 3 were related to the Asian Crisis, 2 were around COVID, 1 from the Great East Japan Earthquake, 1 when Resona Bank was nationalized, 1 post 80s bubble crash, and this month). I’m not entirely certain what this means but such a highly disperse month with such a strong beta move had historically predicated a looming correction. The buyers this month was, again, foreigners while retail, pension plans, and domestic financials were net sellers (with the only other buyer being corporate buybacks).
 
The reason for this dispersion was very clear … “SaaS-pocalypse” as the media is calling it. We saw indiscriminate selling across application software and SaaS businesses across the world after the release of Anthropic’s Claude Cowork which showed the possibility that agentic AI might substitute SaaS solutions. AI changed from a tool to be adopted to a threat that will marginalize workflows. This release came at a time when some US SaaS businesses saw slowing cash-flow growth. Furthermore, refinancing risks were rising given that SaaS debt had low credit ratings, adding to the valuation risk narrative a credit risk element, particularly among US mid-cap software firms. As such, the MSCI World Software & Services index was down -9.1% after falling -13.4% in January. In Japan, the MSCI Japan Software & Services Index was down “only” -5.9% in Jan but accelerated to -14.4% in Feb. Conversely, Japan’s Technology Hardware & Equipment index was up +16.8% (-2.3% in Jan) while the Semiconductors & Semiconductor Equipment Index was up +6.7% (+29.4% in Jan).
 
In our portfolio (currently 12 names), we had one name that was down as much as -18% during the month while we had three that were up over +30%. In our diversified portfolio (currently 40 names), we had an additional name that was down -20% but an additional 2 names that were up over +30% and one other that was up nearly +50%. Now I don’t believe any of them deserved to be up/down that much in a single month, and I therefore did rebalance a little more than normal at around 5~10% per side depending on the portfolio, obviously reducing a little hardware and adding a little software.

​

I expect to do more but the potential extent to how far these software & services stocks can fall is hard to estimate as past parallels are difficult to make. The dot com crash saw many firms disintegrate that had no profit, no revenue, and many with no product. Valuations were based on potential revenue from a total addressable market that had yet to be discovered. Triggered by a rate hike, capital withdrew from the industry as a whole. SaaS-pocalypse are affecting companies with real customers, high recurring revenue, strong growth (both realized and potential), and highly profitable. The crash is causing multiples to contract due to questions about future cash-flow durability. Capital is not being withdrawn but rotated out of software and into Agentic AI (including its infrastructure). I look at a random list of companies that appear to have been particularly hit in the US like DocuSign, Salesforce, and Adobe and interestingly, EPS estimates have been rising despite the stocks falling -30%~-40% this year alone. As such, their multiples look ridiculously low (Disclaimer: I bought what little PA I do via the tax-free NISA program that started a few years ago in a US-listed SaaS-related company after the crash; please note that this is a personal view on this particular stock). But I look to such companies that might guide how far Japan software names might fall. Still, most of our software names, whichever portfolio, are still at or below average position sizes while most of the hardware names are at or above even after the rebalance in February.

​

Now, I argue that Japan and US are vastly different when it comes to SaaS penetration or even IT usage in general. Workflow is very different by company and the systems used are often incompatible and per-seat pricing plans are not common. But that hasn’t prevented a broad range of software solutions, system integrators, and in fact anything that has an IT tilt in their business to be caught up in the negative AI sentiment spillover. I’m hoping the market will remain cautious in order for us to confidently pick up these services growing their bottom line (and often times their top line as well) at double digit growth rates with low probability of true AI disruption and, if anything, explicit AI monetization. In a future monthly, I hope to provide specific examples in our portfolio once we have completed our rebalance.

​

I expected 2026 to be volatile … that appears to have been an understatement. But as I always say, volatility means opportunity.

​

Masaki Gotoh​​​

​

===

​​

Given all that has happened in February and early March, admittedly, I had not the time to think through what to write about this month (and it didn’t help that I was down with the flu for a week as well). So, I chose to skip this month’s random blurb but will resume next month. My apologies.
 ​