Between 30 July and 7 August, 46 Japanese companies worth more than ten billion dollars report. This is peak earnings season for the world's third-largest economy, compressed into a week and a half.
And at its centre sit five companies most Western investors do not know how to read - and without which global trade cannot be understood.
What a sogo shosha is, and why it matters
Japan's five trading houses - Mitsubishi, Mitsui, Itochu, Marubeni and Sumitomo - are not holding companies and not commodity funds. They are the pipe.
They buy natural gas in Australia and sell it in Taiwan. They finance a copper mine in Chile and move the output to China. They own retail chains, salmon farms, power stations, aircraft and shipping fleets. They are the system through which Japan buys and sells the world - and along the way, much of what the world sells to itself.
And why that makes their results a gauge
An ordinary commodity company reports a price. A trading house reports movement - how much flowed, where, at what margin, and at which stage of the supply chain the value was created. When five of them report the same quarter in the same week, the result is a picture of global trade with no American equivalent.
The quarter: five reports, five days
A methodological note before the numbers: the five label the same quarter differently - some call it "Q1 FY2026", others "the year ending March 2027". To avoid confusion, every figure here refers to the three months ended 30 June 2026.
| Trading house | Reported | Revenue | Net profit | Change |
|---|---|---|---|---|
| Sumitomo | 31 July | 1,949.4 | 190.1 | +11.2% |
| Mitsubishi | 3 August | 5,181.0 | 298.5 | +47.0% |
| Itochu | 3 August | 3,875.9 | 293.8 | +3.5% |
| Marubeni | 3 August | 2,609.2 | 186.4 | +20.7% |
| Mitsui | 4 August | 4,347.6 | 294.1 | +53.4% |
In billions of yen
Two first-quarter records in two days. Mitsui declared its highest-ever first quarter with a 53.4% jump, and Mitsubishi beat consensus - ¥298.5 billion against ¥262.2 billion expected.
The engine: commodity prices
The story is uniform across all of them, and it starts in metals and energy.
At Mitsubishi sits the sharpest figure in the group: the mineral resources segment - copper, Australian coking coal, iron ore - posted profit of ¥86.6 billion, up 215%. The company's own bridge is precise: coking coal added ¥19.0 billion to profit, copper ¥20.0 billion, and iron ore subtracted ¥1.0 billion.
And its energy segment jumped 77% - largely from the start-up of its North American LNG business, which contributed ¥31.6 billion against just ¥6.2 billion a year ago.
At Marubeni the same direction: adjusted profit hit a record ¥177.0 billion, within which the resources arm leapt from ¥28.0 billion to ¥51.0 billion - near doubling. Chilean copper and Australian coal are named there too. And its core operating cash flow set a record at ¥249.9 billion, up ¥96.0 billion.
And at Mitsui gross profit rose ¥112.3 billion - from energy, chemicals and the innovation division.
The trap: Itochu
Here you have to read beneath the headline, and anyone who does not will conclude the exact opposite of the truth.
Itochu posted net profit of ¥293.8 billion - up just 3.5%. Against sister companies up 47% and 53%, that looks like a sharp lag.
It is not. Its core profit - profit excluding one-off capital gains - jumped 38% to ¥249.5 billion, a first-quarter record.
The explanation, and a lesson in reading reports
The gap comes entirely from one line: one-off capital gains collapsed from ¥103.0 billion to ¥44.5 billion - a fall of ¥58.5 billion.
Last year Itochu sold large assets and booked enormous gains. This year it sold less. The result: the headline looks flat while the business itself is running 38% ahead.
This is exactly the pattern that has followed us through the earnings season, only in the opposite direction from Amazon: the bottom line stopped describing the business.
The other side of the coin: Mitsui and working capital
And here is the figure no headline carried, and it matters.
Mitsui posted record profit - and negative free cash flow of minus ¥46.4 billion. Its operating cash flow collapsed to just ¥42.3 billion, against ¥262.6 billion a year ago.
The reason is explicit in the report: a working-capital build of ¥271.5 billion, mainly an increase in trade receivables.
This is the direct cost of a commodity boom. When prices rise, the same volume of goods passing through the books requires more cash to finance. Profit is booked immediately; the cash arrives later. Anyone reading only the profit line misses it entirely.
The banks: the rate is finally working
And here a common misconception needs correcting: Japan's three megabanks did not report on the same day. Mizuho reported on 30 July, Sumitomo Mitsui on 31 July, and MUFG on 3 August - a three-day spread across two weeks.
MUFG posted net profit of ¥809.4 billion, up 48.2% - a first-quarter record.
But the line that explains everything is a different one: net interest income reached ¥882.4 billion, up 27.7%.
The numbers proving the reversal works
MUFG's domestic lending rate rose from 1.13% to 1.47% - up 34 basis points. Its deposit rate rose from 0.17% to just 0.30% - 13 basis points.
The spread widened from 0.95% to 1.16%. The bank charges borrowers more and passes only about 38% of the increase to savers.
The context: the Bank of Japan raised its rate to 1.0% on 16 June 2026, the highest since 1995, and held it there on 31 July in an 8-1 vote - where the sole dissenter wanted 1.25%.
This connects directly to what we wrote in the weekly review: the Japanese 30-year bond at 3.975%, after twenty years below 2%. The banks are the direct beneficiaries, and MUFG's report is the first measurement of it.
And Toyota: the number that vanished
Toyota reported today, and it is the most complex report in the group.
- Revenue: ¥13,525.4 billion, +10.4%
- Operating income: ¥1,063.5 billion, down 8.8% - margin fell from 9.5% to 7.9%
- Net income: ¥1,477.0 billion, up 75.6%
How does operating income fall while net income jumps 75%? The answer is below the line: other finance income of ¥850.6 billion against ¥153.7 billion a year ago, and a foreign exchange gain of ¥112.4 billion against a ¥212.4 billion loss. Plus the disposition of Toyota Industries shares.
And North America returned to profit: ¥185.5 billion, a swing of ¥206.6 billion from a ¥21.2 billion loss a year ago. It is the segment's first operating profit since the tariffs took effect.
And what deserves attention precisely because it is absent
Toyota removed the tariff line from its disclosure.
In the same quarter a year ago it published an explicit figure: a tariff impact of minus ¥450.0 billion for the quarter, as a separate line on the summary slide and in the operating-income bridge. For the full year it reported minus ¥1,380.0 billion.
This quarter there is no such line at all - not in the financial summary, not in the presentation, not in the bridge. The revised full-year guidance contains no tariff assumption either.
The only mention is qualitative, from company management: "North America returned to profitability, driven by factors such as the effects of price revisions, foreign exchange effects, and the decrease in U.S. tariffs." Direction only, no magnitude.
This is a change in disclosure, and we note it as such without ascribing motive. But for anyone trying to quantify the cost of tariffs to Japanese industry, that number is simply no longer available.
What else is coming this week
Not yet reported, so there is nothing to say about them yet:
Nippon Steel - which owns U.S. Steel following the acquisition, and is probably the tightest US link any Japanese company has. It reports today after the Tokyo close.
Honda, arriving at its report after a year in which it posted an operating loss of ¥414.3 billion following ¥1,577.8 billion of EV-related write-downs. Excluding those, adjusted operating profit was ¥1,039.3 billion. Its North America plan now rests on combustion and hybrid models after cancelling its EV programmes there.
SoftBank Group on Thursday, Lasertec the same day - and for Lasertec this is a full-year result rather than a quarter, because its fiscal year ends in June. And Recruit Holdings on Friday - which owns Indeed and Glassdoor, making its report one of the most direct reads on the US labour market that exists anywhere.
My Angle
A personal opinion of Ilan Abramov - not advice, not a recommendation
What catches me about this week is not a single number but the consistency.
Five trading houses, five different business models, and all of them pointing at the same thing: metals and energy prices rose, and margins along the chain widened. When Mitsubishi says "coking coal added ¥19 billion and copper ¥20 billion", and Marubeni says "Chilean copper and Australian coal", that is not coincidentally the same answer. It is a measurement of the same reality from two angles.
And two points I carry forward.
The first is Mitsui's working capital. Negative free cash flow alongside record profit is exactly the kind of detail that disappears in headlines. A commodity boom is not only profit - it is also a cash requirement.
The second is the tariff line that vanished at Toyota. I do not know why it was removed, and I do not claim to. But I do know that in the quarter North America returned to profit, the figure that would have let anyone measure that precisely stopped being published. That is worth watching next quarter.
