Gyokuro Time

Whether AI demand is real is a question the market has already answered, and it did not need to wait for SK Hynix’s results. What broke in Seoul on 13 July was not a verdict on that proposition. It was a positioning accident. The Korea Composite Stock Price Index (KOSPI) closed that Monday down 8.95% at 6,806.93, having opened at 7,412 and touched 6,783. At 1:28pm the Korea Exchange tripped its circuit breaker for the seventh time this year and the thirteenth in its history. SK Hynix fell more than 15% to ₩1.845m; Samsung Electronics lost 10.7% to ₩254,000. The flow split cleanly. Foreign and institutional accounts sold a combined ₩3.9tn (about $2.6bn); retail bought ₩3.88tn, almost the mirror image. Individuals were averaging down. ...

When DeepSeek released its low-cost large language model in January 2025, semiconductor shares fell. The logic was simple. If a model can be built cheaply, the expensive GPUs matter less; efficiency rises, so demand falls. That logic has, so far, been wrong. The unit price of inference has dropped by nine-tenths since 2023. Yet the tokens consumed have not fallen but exploded. On Bain & Company’s reckoning, the price per token halved between December 2024 and December 2025 while tokens consumed grew 4.5 times over the same period. Microsoft’s chief executive, Satya Nadella, greeted DeepSeek’s release with the line “Jevons paradox strikes again”. The nineteenth-century economist William Stanley Jevons found that as the steam engine grew more efficient, coal consumption rose rather than fell. A resource that gets cheaper is not used less; it opens new uses and its total climbs. ...

On 17 July the Nikkei 225 closed at 64,141, down 2,694 points on the day. That is a fall of 4.03%. As a one-day drop it ranks fifth on record; as a week it was the largest ever, at 4,416 points. Semiconductors and the wider AI complex led the way down: Kioxia Holdings fell 16.1%, Ibiden 9.9%, Tokyo Electron 8.2% and Advantest 7.2%. The headlines reached for a warning about the AI growth story. Line the day’s moves up in a row, though, and the story that was supposedly under threat is nowhere to be seen. What shows up is duller than that. ...

In the second quarter of 2026 Taiwan Semiconductor Manufacturing Company spent $15.7bn on plant and equipment. Add the first quarter and the half-year total reaches $26.8bn. On the same day it reported those figures, the company raised its full-year capital budget to a range of $60bn-$64bn, up from the $52bn-$56bn it had guided in January. Management gave two reasons: demand that keeps climbing, and tools now bought at inflation-era prices. Free cash flow fell by NT$60.8bn over the quarter, which the company attributed to capital spending outrunning the growth in operating cash. The quarterly report sets it all out. ...

The lobster is sometimes said to be immortal. It shows almost no sign of senescence: it does not weaken or lose fertility with age, and its cells keep dividing as though the years did not count (Smithsonian). But not ageing is not the same as not dying. The lobster dies. It dies not of old age but of the very mechanism of its growth. A lobster can grow only by shedding its whole shell in a moult, and each moult costs exponentially more energy as the animal gets larger. Between 10% and 15% of older lobsters die each year from a failed moult, and in time they stop moulting at all to conserve energy (Smithsonian, citing Maine’s Department of Marine Resources). The old shell then wears, bacteria get in, and scar tissue fixes the body to the shell. Trapped in the shell it will not shed, the animal rots. It dies because it stopped shedding, not because its time was up. ...

The speculative yen short that had built to a 19-year high fell for the first time in the week to 7 July. On the Commodity Futures Trading Commission’s positioning data, the leveraged-fund net short in the yen, futures and options combined, went from −137,828 contracts on 30 June to −104,231, a cover of 33,597 in a week. The last time the book ran heavier was the first half of 2007. Yet the yen barely moved. ...

As interest rates rise, firms buckle. Put July’s run of regional-bank losses beside the collapse of Zentoshin, an Osaka payment agent, and the story writes itself. It is half wrong. The Bank of Japan did raise its policy rate to 1.00% in June, the highest since 1995, but stripped of inflation the real rate is still deeply negative (Dai-ichi Life Research Institute). A rate of 1% is not, in itself, crushing borrowers. What changed is not the level of rates but the machinery that ran beneath them. ...

Few people had heard of the company that went bankrupt on 6 July. Zentoshin, a credit-card settlement agent in Osaka, was no household name. Yet it sat beneath the cash flow of restaurants and the night-time economy across Japan, working as their plumbing. Its liabilities came to about ¥125.9bn (roughly $790m), the largest corporate failure of the year and the first above ¥100bn since Drone Net in December 2025 (TDB, TSR). Its collapse pulled a scatter of regional banks down with it. ...

Markets have settled on a number. America’s tariff on most of its trading partners sits at 15%, and the figure has barely twitched in a year. That stability is the trap. The 15% has held while the legal ground beneath it was struck down, replaced, struck down again and rebuilt under a third statute. The rate is the one thing not moving. Everything underneath it is. This is the part the consensus underprices. A tariff is usually read as a price: a level that feeds margins, inflation and the terms of trade. A tariff imposed by presidential power is also a legal instrument with a shelf life, and the shelf life is where the information sits. There is a larger contest inside it, too. A great power is holding a tariff wall in place by hopping from one emergency power to the next as the courts strike each one down: a quiet test of how much the executive can tax without Congress, refought every few months. The market reads that test, imperfectly, as a number. Watch the rate and you learn little. Watch which statute is carrying it, and when that statute breaks, and you are reading a calendar of discrete events, each one able to move the carry trade, the business of borrowing where money is cheap to invest where it is dear, more than the tariff itself ever will. ...

In 2025 Japan ran a current-account surplus of ¥31.9trn, a record and the second in a row. Yet the yen has not stopped falling. Even after the Bank of Japan raised rates in June 2026, the currency trades around ¥161 to the dollar, close to its weakest since 1986. A record creditor’s currency sits at a four-decade low. Unless that contradiction can be resolved, the question every market participant is asking has no answer: where does the yen go next? ...