Compare a sector from its filings with Claude and Kabuken MCP: Japanese automakers
Most sector comparisons start from a data vendor's screen. Someone has already decided what "revenue" means for each company and which year's restated figure to show. When a number looks odd, you cannot open the source.
This walkthrough does it the other way round. We compare four Japanese automakers using only their annual securities reports (有価証券報告書) filed on EDINET, the Financial Services Agency's disclosure system, read through Claude with the Kabuken MCP connector. Every figure in this article carries the EDINET document ID of the filing it came from, so you can open the filing and check it.
The four companies are Toyota Motor (EDINET code E02144), Honda Motor (E02166), Suzuki Motor (E02167) and SUBARU (E02152). All four report under IFRS for the fiscal year ended 31 March 2026, so the headline table compares like with like. Amounts are in millions of yen, as filed.
Kabuken (kabuken.intellics.ai) reads the XBRL data in EDINET filings and hands it to an AI assistant with the source filing attached to each number. It provides no stock prices and no investment advice.
Before you start
Connect Kabuken to Claude once: in Claude on the web, go to Customize, then Connectors, add a custom connector with the URL https://kabuken.intellics.ai/mcp, sign in to Kabuken and select Authorize, then turn Kabuken on from the "+" menu in the chat box. The full guide, including Claude Code and the desktop app, is here: https://kabuken.intellics.ai/docs/en/clients/claude-ai/
Before each tool call, Claude shows a permission panel naming the Kabuken tool it is about to use. Approve it once or always.
Step 1: find the filings
List the latest annual securities reports (有価証券報告書) for Toyota, Honda,
Suzuki and SUBARU, with the document ID, fiscal period and filing date.
This request maps to search_companies (one call covering the four names), then list_documents once per company. The list includes every EDINET filing by or about the company (extraordinary reports, buyback reports, large-shareholding reports), so the annual report is the row whose doc_type_code is 120. Trimmed output:
| Company | Fiscal year | Period | Filed | doc_id |
|---|---|---|---|---|
| Toyota Motor | 第122期 | 2025-04-01 to 2026-03-31 | 2026-06-10 | S100Y8NY |
| Honda Motor | 第102期 | 2025-04-01 to 2026-03-31 | 2026-06-18 | S100YDD9 |
| Suzuki Motor | 第160期 | 2025-04-01 to 2026-03-31 | 2026-06-23 | S100YFG2 |
| SUBARU | 第95期 | 2025-04-01 to 2026-03-31 | 2026-06-22 | S100YFKZ |
These four IDs are the anchors for everything that follows. Suzuki filed an amendment S100YYB1 on 2026-08-25 that re-attaches the report's inline XBRL; the summary figures used here (revenue, profit before tax, profit attributable to owners, equity, total assets) are the same in both filings, and this article cites the original S100YFG2.
Step 2: the headline table
Compare Toyota (E02144), Honda (E02166), Suzuki (E02167) and SUBARU (E02152)
on revenue, operating income, net income, total assets and equity from their
latest annual reports. Show the document ID and accounting standard for each.
This maps to compare_companies. Each metric in the response carries the XBRL concept it was read from (concept_id), the period and the consolidation basis, and the response states each company's accounting_standards. All four say IFRS.
FY2026-03, consolidated, IFRS, JPY million:
| Toyota | Honda | Suzuki | SUBARU | |
|---|---|---|---|---|
doc_id |
S100Y8NY | S100YDD9 | S100YFG2 | S100YFKZ |
| Revenue | 50,684,952 | 21,796,610 | 6,292,967 | 4,784,965 |
| Operating income | 3,766,216 | -414,346 | 622,909 | 40,120 |
| Net income | 3,848,098 | -423,941 | 439,267 | 90,842 |
| Total assets | 105,522,331 | 33,509,285 | 6,636,815 | 5,492,301 |
| Equity | 39,918,854 | 11,817,512 | 3,382,083 | 2,780,407 |
| Operating margin (computed) | 7.43% | -1.90% | 9.90% | 0.84% |
| Equity ratio (computed) | 37.8% | 35.3% | 51.0% | 50.6% |
Three definitions matter here, and the tool response makes each of them visible:
- "Net income" is profit attributable to owners of the parent (concept
ProfitLossAttributableToOwnersOfParentIFRSSummaryOfBusinessResults). For Toyota that is 3,848,098, not the total profit for the year of 3,985,761, which includes non-controlling interests. - "Equity" is equity attributable to owners of the parent, on the same basis.
- Toyota's top line is 営業収益 (operating revenues, concept
TotalNetRevenuesIFRS). It is the sum of sales of products (45,865,949) and revenue from financing operations (4,819,003). Honda, Suzuki and SUBARU reportRevenueIFRS. The compare tool labels all four as "Net sales / Revenue"; the concept ID tells you which one you are looking at.
The two computed rows are ours, not the filing's: operating margin is operating income divided by revenue, and the equity ratio is equity attributable to owners divided by total assets. The response includes an operating_margin_denominator field for exactly this purpose, and its attribution block adds the line "Figures are calculated by Kabuken from EDINET filings, not reported directly by EDINET."
A note on reading the table: Toyota and Honda both consolidate large financial-services arms, which carry loan and lease receivables on the balance sheet. That is one reason their equity ratios sit well below Suzuki's and SUBARU's. Step 4 shows the size of those arms.
Step 3: three years, each from its own filing
Show the last three fiscal years of revenue and operating income for each of
the four companies, taking each year from that year's own annual report, and
mark the accounting standard for each year.
This maps to get_company_financials with period_type: ANNUAL and periods: 3. Each period in the response carries its own doc_id, so a row is the figure as first filed, not a later restatement.
JPY million:
| Company | FY2024-03 | FY2025-03 | FY2026-03 |
|---|---|---|---|
| Toyota revenue | 45,095,325 (S100TR7I) | 48,036,704 (S100VWVY) | 50,684,952 (S100Y8NY) |
| Toyota operating income | 5,352,934 | 4,795,586 | 3,766,216 |
| Honda revenue | 20,428,802 (S100TNE0) | 21,688,767 (S100VYOD) | 21,796,610 (S100YDD9) |
| Honda operating income | 1,381,977 | 1,213,486 | -414,346 |
| Suzuki revenue | 5,374,255 (S100TRQC, J-GAAP) | 5,825,161 (S100W4MT, IFRS) | 6,292,967 (S100YFG2, IFRS) |
| Suzuki operating income | 465,563 (J-GAAP) | 642,851 (IFRS) | 622,909 (IFRS) |
| SUBARU revenue | 4,702,947 (S100TO9E) | 4,685,763 (S100W1AE) | 4,784,965 (S100YFKZ) |
| SUBARU operating income | 468,198 | 405,308 | 40,120 |
Toyota, Honda and SUBARU report under IFRS in all three years. Suzuki's FY2024-03 report is under Japanese GAAP (concepts jppfs_cor:NetSales and jppfs_cor:OperatingIncome); its later reports are IFRS. The response states the standard per period (accounting_standards), which is how the J-GAAP row is labelled above. Net sales under J-GAAP and revenue under IFRS are not the same quantity, and operating income is defined differently too, so we do not compute a change across that boundary. For the other three companies the direction is visible without arithmetic: revenue rose at Toyota and Honda over the three filings, while operating income fell at Toyota in each year and turned negative at Honda in FY2026-03. SUBARU later amended its FY2025-03 report (訂正有価証券報告書 S100YF9D, 2026-06-22); the FY2025-03 figures above are from the original S100W1AE.
Step 4: what each company is made of
Break down each company's latest annual report by business segment:
revenue from external customers and operating profit per segment, with the
Japanese and English segment names.
This maps to get_segment_breakdown on each doc_id. The response comes from the XBRL segment dimensions of the filing (axis jpcrp_cor:OperatingSegmentsAxis), so the Japanese names below are the companies' own.
Segment revenue is revenue from external customers (RevenueFromExternalCustomersIFRS, or Toyota's OperatingRevenueFromExternalCustomersIFRS). Operating profit is OperatingProfitLossIFRS per segment. FY2026-03, JPY million:
Toyota (S100Y8NY)
| Segment | 日本語 | Revenue (external) | Share (computed) | Operating profit |
|---|---|---|---|---|
| Automotive | 自動車 | 45,201,924 | 89.18% | 2,777,049 |
| Financial Services | 金融 | 4,819,003 | 9.51% | 851,722 |
| Other | その他 | 664,026 | 1.31% | 132,079 |
| Corporate expenses and elimination | 消去又は全社 | 5,366 |
Toyota's financial-services segment is 9.51% of revenue and 22.61% of operating income (851,722 of 3,766,216, computed). Its external revenue, 4,819,003, is the same figure that appears as "revenue from financing operations" in the income statement in step 2.
Honda (S100YDD9)
| Segment | 日本語 | Revenue (external) | Share (computed) | Operating profit |
|---|---|---|---|---|
| Automobile business | 四輪事業 | 13,863,362 | 63.60% | -1,411,140 |
| Motorcycle business | 二輪事業 | 4,018,837 | 18.44% | 731,926 |
| Financial services business | 金融サービス事業 | 3,529,484 | 16.19% | 275,532 |
| Power product and other businesses | パワープロダクツ事業及びその他の事業 | 384,927 | 1.77% | -10,664 |
Honda's four segments sum exactly to the consolidated figures: 21,796,610 of revenue and -414,346 of operating profit. The automobile business reported an operating loss of 1,411,140; motorcycles reported an operating profit of 731,926 and financial services 275,532. The same segment note carries impairment losses (ImpairmentLossesPLIFRS, 692,006 for the segments in total, of which 605,150 in the automobile business), capital expenditure and depreciation, which you can ask for in the next prompt.
Suzuki (S100YFG2)
| Segment | 日本語 | Revenue (external) | Share (computed) | Operating profit |
|---|---|---|---|---|
| Automobile business | 四輪事業 | 5,706,420 | 90.68% | 547,632 |
| Motorcycle business | 二輪事業 | 454,488 | 7.22% | 44,770 |
| Marine business | マリン事業 | 119,456 | 1.90% | 26,605 |
| Other business | その他事業 | 12,601 | 0.20% | 3,900 |
SUBARU (S100YFKZ)
| Segment | 日本語 | Revenue (external) | Share (computed) | Operating profit |
|---|---|---|---|---|
| Automobiles | 自動車 | 4,638,342 | 96.94% | 32,086 |
| Aerospace | 航空宇宙 | 141,667 | 2.96% | 3,500 |
| Other | その他 | 4,956 | 0.10% | 3,616 |
| Corporate expenses and elimination | 消去又は全社 | 918 |
Segment figures are rounded to millions in the filings, so a column can differ from the consolidated total by a few million yen (Toyota's three revenue segments sum to 50,684,953 against a reported 50,684,952). The shares are computed against the consolidated revenue in step 2.
Read across the four tables and the sector stops looking uniform: two companies with financial-services arms of 9.51% and 16.19% of revenue, one with a motorcycle business near a fifth of revenue, one with an aerospace segment, and two that are more than nine-tenths automobiles. None of this is in the headline table.
Step 5 (optional): who has filed a large-shareholding report
Show SUBARU's major shareholders from large-shareholding reports, with each
holder's percentage, report date and document ID.
This maps to get_major_shareholders (a looser wording such as "which reports have been filed recently" can be answered from the Step 1 list without this tool, so ask for the percentages). The rows are large-shareholding reports (大量保有報告書) and their change reports (変更報告書), which holders file when they cross the 5% threshold or change their position. They are not the shareholder register printed in the annual report, and each row is one report, dated. Trimmed:
| Filer | Holding | Report date | Type | doc_id |
|---|---|---|---|---|
| 野村アセットマネジメント株式会社 | 5.31% | 2026-07-07 | new | S100YPB1 |
| ノルウェー銀行 (Norges Bank) | 4.33% | 2026-02-18 | change | S100XM2E |
| ノルウェー銀行 (Norges Bank) | 5.04% | 2025-12-18 | new | S100XAW3 |
Each row's purpose field repeats the holding purpose as filed (for Norges Bank, 純投資, pure investment).
What this article does not do
- No stock prices, valuation ratios or market data. EDINET filings do not contain them and Kabuken does not add them.
- No forecasts, no view on which company is "better", no buy or sell language. The tables describe what was filed.
- No normalisation across accounting standards. Suzuki's J-GAAP year is labelled and left alone.
- Segment names are the companies' own reportable segments. "Automobile business" at Honda and "Automotive" at Toyota are not defined identically.
- Claude writes the prose around the numbers. Check its sentences against the tool output and the filing, as you would with any analyst's note.
The one thing to take away
Every number above has a doc_id. S100Y8NY is Toyota's 第122期 annual securities report, filed 2026-06-10; open it on EDINET and the figures are on the face of the statements and in the segment note. That traceability, not the chat interface, is what makes the comparison usable in a report of your own.
Attribution
Source: EDINET (Financial Services Agency, Japan), Public Data License 1.0 (https://www.digital.go.jp/resources/open_data/public_data_license_v1.0). Processed by Kabuken. This content is based on EDINET filings and has been edited/processed by Kabuken; it is not presented as unprocessed government data. Rows marked "computed" are calculated by the author from the filed figures shown in the same table. Segment axis and member labels include EDINET Taxonomy content, which is excluded from PDL 1.0 and separately licensed (https://www.fsa.go.jp/search/EDINET_Taxonomy_Legal_Statement.html). © Copyright 2014 Financial Services Agency, The Japanese Government.
- Kabuken: kabuken.intellics.ai
- Connect Kabuken to Claude: https://kabuken.intellics.ai/docs/en/clients/claude-ai/
- EDINET: https://disclosure2.edinet-fsa.go.jp/