Around the 8th century BCE, Bian He, a jeweller, found a stone he knew from experience to be priceless jade. He presented it to King Li, who took him for a deceiver and had him cast out. He brought it to the next king, King Wu, who dismissed him the same way. Twice humiliated, he sat weeping at the foot of Jingshan Mountain — until King Wen finally asked why he wept. "I do not grieve for myself," Bian He replied, "but that something of true worth should be mistaken for nothing at all." The stone was cut open and found to be jade, pure white and flawless. The story originates in ancient China, but passed into Japanese culture along with the characters, classics, and moral teachings that shaped it — and it endures, in both traditions, as a parable of hidden treasure: that what appears rough and worthless may contain something extraordinary within. Which feels relevant for now.
While the market has spent the spring transfixed by AI infrastructure and semiconductors — leading the Nikkei to behave like a leveraged version of the US Semiconductor Index, and the world index to look little different — we have been doing something rather different. We have been gathering our unpolished stones: businesses that are cheap, cash-generative, and in many cases entering earnings upturns that the market has not yet noticed. Not because we are contrarian for the sake of it, but because we have done the work, cut the stone open, and know what is inside.
Our portfolio is built around three kinds of hidden treasure. None of which require AI to keep spending at current rates. All of them work in a rising-rate, stronger-yen environment. And collectively, they offer something that has become genuinely rare: diversification.
Under Earners
The first kind of business has been structurally suppressed — by management complacency, by industry overcapacity, and by a capital-allocation culture that historically prized scale over returns. To the court craftsmen, the stone looks unremarkable. The earnings look pedestrian. But that is a reflection of how the business has been run, not of what lies inside. And that is changing.
Sumitomo Chemical is a case in point. Japanese petrochemicals have endured a long and painful rationalisation — too many players, too much capacity, in a world that has structurally shifted beneath it. But that very process is now producing a smaller, more focused industry with better pricing power. We are not buying Sumitomo Chemical on a view about global chemicals demand. We are buying it because the earnings power of a rationalised business is materially higher than today's numbers imply — and the valuation reflects none of it.
Air Water tells the same story in a different industry. Industrial gases, the world over, is a superb business: annuity-like economics, enormous switching costs, and the high returns that follow. Linde earns operating margins of around 25%. Air Water, with the same underlying economics, earns just under 7% — the difference: a lack of focus and strategic drift. The refocusing is now underway. The gap between current earnings and normalised earnings power is wide.
These are companies for which the path from current earnings to normal earnings is visible, dateable, and most importantly, not yet in the price.
Mispriced Cashflows
The second kind of business, mispriced cashflows, is where the cognitive dissonance between what we own and what the market is chasing becomes most stark.
We own a group of companies growing at +10% or more a year, many generating free cash flow yields in excess of 20% at current prices. A 20% free cash flow yield, on growing businesses, at a time when certain technology names trade on implied free cash flow yields, based on 2029 forecasts, of 1% to 2%. The market is paying almost anything for a story, and almost nothing for cash in hand.
Park24, the car-park operator, generates recurring, predictable cash flows from an asset base with genuine pricing power and high barriers to entry. Tsubakimoto Chain supplies precision motion components into long replacement cycles, with strong aftermarket economics. Neither mentions AI in its investor presentations. Neither has been written up by the sell-side desks currently tripping over one another to cover Kioxia and Advantest. That is part of why they are cheap — and entirely why they are interesting.
An ordinary rock, to the king's craftsmen. To anyone willing to cut it open, is something rather finer.
Overlooked Assets
The third kind of business is the most literal expression of Bian He's story: companies where the price on the screen bears little relation to the value sitting on the balance sheet, and where a catalyst now exists to bring the two together.
The regional banks — Kyoto Financial, Shiga Bank, Hachijuni — are the clearest example. Decades of cross-shareholding relationships have left them holding enormous portfolios of equity investments. In several cases, the value of the equity investments alone exceeds the market capitalisation of the bank. Put plainly: you are handed the equity portfolio and pay almost nothing for the banking business attached to it. The Tokyo Stock Exchange's pressure to unwind cross-shareholdings and improve capital efficiency is the catalyst. It is slow. But it is happening.
NS United Kaiun operates a fleet of bulk carriers and specialised vessels whose replacement value comfortably exceeds the company’s market capitalisation. It holds net cash. It pays a substantial dividend. There is nothing complicated here — only a balance sheet that rewards the reading of it. Arealink, in real-estate storage, has genuine scarcity economics and recurring revenue. Katakura Industries sits on a collection of real estate, securities, and operating businesses worth considerably more than their whole core business.
What strikes us, looking across these holdings, is how much of their value is already visible before the operating business is considered at all. Latent assets — investment securities, land, net cash — represent around 43% of the portfolio’s market capitalisation. For every pound invested, roughly 43p is already backed by identifiable assets before anyone has placed a value on the operating business at all. These are not speculative positions — they are compounding engines that need no external validation from a macro theme. The downside is anchored, and the upside has simply not yet been recognised.
Beauty is in the eye of the beholder: what is just a rough rock to some, when examined, is so much more. When the assets on the balance sheet are worth more than the entire company, the downside is structurally limited — and that margin of safety is the whole point.
A word on the stones everyone else is holding
We are not dismissive of AI. The technology is real, the demand is real, and the companies riding hyper-scaler capital expenditure have genuine orders today. The question we keep returning to is simple: how long can this level of spending be sustained without a commensurate and visible return?
The numbers are becoming hard to ignore. Across the hyper-scalers, capital expenditure now exceeds free cash flow. The gap is being met by debt, and increasingly by equity. Every new data centre demands fresh capital, and the returns on that capital have not been demonstrated at scale. The moat is unclear: the leading model changes with each release, and the cost of staying at the frontier is extraordinary.
The market behaviour, fixated on a single theme, violent rotation within it, wild intraday ranges, and valuation treated as an afterthought — is the signature of late-stage momentum. That does not mean it ends tomorrow; these things run longer than they ought to. But we are well into the process, and demand is now being met by a rising wall of new equity supply. Elevated valuations, unproven economics, and rising supply - that combination rarely ends well, even when the underlying technology is genuinely transformative.
These, one might say, are the stones the whole court is admiring at once. We would rather hold the ones nobody has looked at properly.
The Point of the Story
Here is what Bian He understood, and what the first two kings did not. The value of a thing and the recognition of its value are two separate events — sometimes separated by years, and by no small amount of patience. The jade was always jade. The stone never changed. What changed was that someone, eventually, looked properly. The market does not always share the temperament. We do.
Sitting across our three buckets today, I see a portfolio that is cheap, diversified, and well-positioned for an environment quite different from the last eight weeks. The under-earners are entering upturns not yet in the numbers. The cashflow businesses are compounding quietly, at yields that look faintly ridiculous beside the momentum trade. The asset-rich names carry a floor beneath them.
At some point, the market, stops looking with disdain at the stone in the king's hand, picks up ours, and cuts it open.
We think that moment is closer than it looks.
250626.280
