AI Impact on Japan's Sogo Shosha — The End of "Information Asymmetry" and the Future of General and Specialty Trading Houses

Itochu ~¥900.3 billion. Mitsui ~¥834.0 billion. Mitsubishi Corp ~¥800.5 billion. The five Japanese sogo shosha (general trading houses) again posted net profits in the ¥800–900 billion range for FY2025 (the year to March 2026), and Berkshire Hathaway holds roughly 10% of each of the five (9.7–10.8% at the end of 2025). Yet underneath that record, a structural shift is shaking the sogo shosha's core business model. On May 19, 2026, the Policy Research Council of Japan's ruling LDP approved a proposal on "Next-Generation AI & On-Chain Finance" — linking AI-agent transactions to the flow of goods and automating their settlement on blockchain. The proposal never mentions trading houses, but the transaction and settlement work at the core of the sogo shosha is now within the reach of national policy.

Up front: "sogo shosha collapse soon" is hype; "half the trading-house jobs disappear" is fact. Brief background for non-Japanese readers: a sogo shosha is a uniquely Japanese conglomerate that combines functions performed in the US by importers, exporters, freight forwarders, banks, law firms, and consultants — all under one roof. Their historic moat is "information asymmetry": monopolizing intelligence on overseas resources, markets, and contacts. But Bloomberg, Reuters, SaaS, generative AI, satellite imagery have built "information infrastructure anyone can access," and the moat is vanishing fast. The shift to "investment company" and "downstream consumer business" has been underway for 30 years — and Itochu taking the No. 1 spot in FY2025 net profit (reported in May 2026) is the clearest signal of the strength of "downstream + AI".

Personal take up front: "I got a sogo shosha offer — my career is set!" is the single biggest illusion of 2026 and beyond. The salary band (¥15M at junior, ¥20M+ by mid-40s) holds, but about half of the typical "shosha-man" workload — intelligence gathering, document prep, credit checks, logistics coordination, FX hedging operations — moves to AI. What remains for humans is "on-the-ground political risk judgment, large M&A, long-term investment, relationship-capital management" — concentrated in the senior tier. New grads through early 30s have only one play: "use AI hard enough to outperform an average senior." This is the same white-collar elimination / seniors-vs-juniors structure, sharpest at the sogo shosha. This article maps the historic moat, AI's four impact areas, the Big Five strategies side-by-side, the May 2026 policy shock, survival strategies, and the shosha-man career map — all grounded in May 2026 data.

1. The "AI Wave" Hits Japan's ¥900-Billion-Profit Trading Houses

Japan's five sogo shosha (Mitsubishi Corp, Mitsui & Co, Itochu, Sumitomo Corp, Marubeni) again posted strong FY2025 results (the year to March 2026): Itochu ¥900.3B (a record), Mitsui ¥834.0B, Mitsubishi ¥800.5B, Sumitomo and Marubeni in the ¥500–600B range. Resource-heavy Mitsubishi and Mitsui saw profits fall, so the top spot in net profit passed to Itochu (in FY2024 Mitsubishi led with ¥950.7B). Berkshire Hathaway announced stakes of just over 5% in each in August 2020, kept buying, and held 9.7–10.8% at the end of 2025 (Mitsubishi 10.8%, Mitsui 10.4%, Itochu 10.1%, Marubeni 9.8%, Sumitomo 9.7%). Its shareholder letter ranks the Japanese holdings as comparable in importance to its major US holdings.

"Record results" and "structural shift" coexist. As of May 2026 three waves are hitting the industry simultaneously: ① the end of information asymmetry (AI × SaaS), ② national-policy-driven transaction automation (the LDP proposal approved on May 19), ③ widening intra-industry gap (Itochu No. 1 in FY2025 net profit). The sogo shosha's core business model — "one-stop information, credit, and logistics" — faces "disintermediation" pressure from AI + blockchain + SaaS.

Understanding the business structure reveals the impact zones. Japanese trading houses cover both "upstream" (resources, energy) and "downstream" (food, convenience stores, housing) — that combination is the distinctive sogo shosha model. Upstream is investment exposed to commodity prices and political risk; AI impact is limited (geology, country judgment remain human). Downstream is consumer-facing; AI / data directly hits it. Itochu is leaning into downstream; Mitsui is doubling down on upstream — the strategies diverge. As covered in white-collar elimination, the "industry × AI impact" mismatch shows up sharply within trading houses too.

2. The End of "Information Asymmetry" — Sogo Shosha's Historic Moat

The sogo shosha earned for over a century on "information asymmetry." Specifically: ① overseas resource location, price, logistics intelligence; ② emerging-market political, regulatory, network intelligence; ③ international finance and FX hedging know-how; ④ B2B counterparty credit intelligence. Because "only the shosha had these," the "middleman margin" worked.

By the late 2020s, all four became "anyone-can-access". ① Resources: Bloomberg / Refinitiv / CME real-time prices, satellite imagery (Planet Labs et al.) monitors ports and mines. ② Emerging markets: Foreign Policy, Stratfor, Eurasia Group analyses sell for tens of thousands of yen monthly; generative AI (Perplexity, ChatGPT) instantly synthesizes them. ③ FX hedging: banks deliver direct, fintech (Wise, Airwallex) brings it to small businesses. ④ Credit: D&B, Experian, plus AI credit scoring (CB Insights and similar) standardize the playing field.

As a result, half or more of the shosha "middleman" model lost its economic logic in 2026. The shosha saw this coming as early as the 1990s and pivoted to "investment company" mode — Mitsubishi's Lawson and KFC, Itochu's FamilyMart and Thai CP partnership, Mitsui's IHH Healthcare (Malaysia's largest hospital group). The "trading → investing" shift is essentially done, but AI is now starting to compress the "investee operations" work too.

3. Four Business Areas Hit by AI

The shosha's work mapped to "AI impact × residual value" falls into four areas. This is our own reading of public information; the percentages are rough guides, not figures from any specific study.

The pivotal data point: AREA 1 (trade execution) accounts for 50–60% of typical shosha-man work. If 70% of it is automated, that is roughly 35–40% of all shosha-man work. The impact is enormous, and each firm now debates "headcount cuts vs profit-per-person growth." Mitsubishi tilts toward the "business-company" path (preserve headcount, raise per-capita productivity); Itochu tilts toward "AI-driven slimming + downstream expansion." The strategic split will decide 2030s performance.

4. AI/DX Strategy of the Big Five — Itochu Leads, Mitsubishi Drifts

Big-Five AI/DX strategy side-by-side as of May 2026:

The standout: Diamond Online's report (April 2, 2026) that words such as "DX" and "digital" have dropped sharply in Mitsubishi Corp's recent integrated reports. Two possible readings: "DX has become a stale term" or "a sign of strategic drift." Either way, the fact that Itochu took No. 1 in FY2025 net profit in the same period is heavy. The industry has shifted from "shosha that talk DX" to "shosha that embed AI in operations" — from language to execution.

Personal call: Itochu's combination — "Silicon Valley investment vehicle + all-hands AI training + AI applied to downstream businesses (FamilyMart, Yanase)" — is becoming the 2026 standard model for sogo shosha. Mitsubishi and Mitsui will sustain profits on resource deals, but on the clarity of "the next decade's growth engine," Itochu is a step ahead.

5. May 2026: The LDP's "Next-Generation AI & On-Chain Finance" Proposal Shock

On May 19, 2026, the ruling LDP's Policy Research Council approved a proposal drafted by the "Next-Generation AI & On-Chain Finance" project team under its Digital Society Promotion Headquarters (LDP announcement, in Japanese). The proposal itself does not mention trading houses, but its content could hit the sogo shosha business directly. The proposal has two core points:

① AI identifies and executes commercial transactions — combined with smart contracts, the "price negotiation, contracting, fulfillment" of trade transactions runs autonomously via AI. The sogo shosha's "trade execution" work is being automated at the level of national policy.

② Blockchain handles settlement and reconciliation — L/Cs, trade finance, and FX settlement get processed via stablecoins and tokenized deposits. Disintermediation pressure on the "financial functions" the sogo shosha historically owned.

If the proposal is fully realized, this stacks on top of 70% AREA 1 automation and adds significant compression to AREA 2 (finance / logistics coordination) — a scenario where "half the sogo shosha workload disappears via national policy." Full implementation will take 5–10 years, but the trading houses have to move now, in the preparation phase — which is why Itochu's "downstream-first" path is the right call.

6. Survival Strategies — Investment Holding, Downstream, AI-Native

Three strategies define sogo shosha survival into the 2030s: ① investment-holding company, ② downstream expansion, ③ AI-native organization.

If sogo shosha morph from "middlemen" into "investment fund + business operator + AI-leveraged" hybrids, there's a clear path through the 2030s. But the old "shosha-man image" (overseas posting, risk-taking, entertainment-driven relationships) will shrink — replaced by a new shosha-man defined by "AI operator + investment judgment + industry-deep expertise."

7. Shosha-Man Career Map — Who Vanishes, Who Stays

Three career layers for shosha-men, with sharply different AI impact:

New grad / early-career advice: "I got a sogo shosha offer — career set!" is the biggest illusion of 2026. Compensation holds (¥8M new grad → ¥15M in 30s → ¥20M+ in 40s), but "junior headcount slots are likely to shrink." The survival play: "be the one who uses AI best among your cohort," "go deep into a specific industry / region," "build relationship capital that connects directly to GM-level executives" — run all three in parallel through your 20s and 30s. The "I-shaped career" and "relationship-capital investment" from jobs that survive the AI era apply directly to shosha-men.

8. Specialty vs General Trading Houses — Gap Widens

Beyond the five sogo shosha, Japan has many specialty trading houses (steel, textiles, food, chemicals, etc.). Under AI pressure, the "sogo shosha advantage, specialty disadvantage" gap is likely to widen.

Reasons: ① The sogo shosha diversify across business portfolios and can spread AI / policy risk; specialty trading houses, concentrated in a single industry, have nowhere to go if their industry declines. ② Sogo shosha have the DX investment budget (¥100B+ over five years); specialty firms have limited capital and lag on AI investment. ③ Specialty trading houses' core work (information brokerage, logistics coordination) IS trade execution — exactly the AREA AI hits hardest.

Highest-risk specialty segments: textile trading houses (info-asymmetry-centric), steel trading houses (price-brokerage-centric), mid-tier food trading houses (logistics-coordination-centric). Specialty firms that build their own brands, develop products, or broker specialized technology survive — e.g. Itochu Marubeni Steel (IMM), Metal One, JFE Shoji, which command sogo-shosha-class strategic flexibility. "Polarization within specialty trading houses" is another key 2026–2030 trend.

Summary

The Japanese trading-house industry has "record profits and structural shift in parallel" in 2026. The historic moat — "information asymmetry" — is vanishing as Bloomberg, SaaS, generative AI, and satellite imagery commoditize intelligence. The "Next-Generation AI & On-Chain Finance" proposal approved by the LDP on May 19, 2026 puts the automation of transactions and settlement, which overlaps with core sogo shosha work, within the reach of national policy. Trade execution, about half of shosha-man work, could be 70% AI-automated by 2030 (roughly 35–40% of all the work).

Big Five strategies have polarized: Itochu (downstream × AI × Silicon Valley investment) took No. 1 in FY2025 net profit; Mitsubishi, reported to have sharply cut "DX" from its integrated reports, runs a business-company path; Mitsui doubles down on resources and smart cities; Sumitomo / Marubeni ride finance, data, and logistics. The three survival strategies: ① investment-holding (Berkshire-style conglomerate), ② downstream expansion, ③ AI-native organization. Only the sogo shosha that commit seriously to all three win the 2030s.

Shosha-man careers split sharply into three layers: new grad – early 30s = high risk (70% of work disappears), late 30s – 40s = AI-operator mandatory, GM+ = market value rises (busy work cleared, focus on judgment). "I'll be set with a sogo shosha offer" is the biggest 2026-onward illusion — assume it, and play AI fluency × industry depth × relationship capital through your 20s and 30s. Specialty trading houses face widening "sogo advantage, specialty disadvantage" as capital and diversification decide outcomes.

Related: white-collar elimination, future of sales, seniors-vs-juniors, jobs that survive AI, 15 jobs at risk.

FAQ

Q. I'm a student with a sogo shosha offer — should I accept?
A. For a Big Five sogo shosha — yes, but prioritize "AI fluency" above all. Total value (comp, career, international exposure) remains high. But unless you "out-AI 100 of your cohort within 3 years," the path to senior tier narrows quickly. Touch Cursor, Claude Code, and industry AI tools before you start.

Q. I've been at a trading house 10 years — should I switch?
A. If you're late 30s, push for an internal "AI initiative" move first. AI fluency + 10 years of industry knowledge is scarce internally. If switching, targets are "industry × AI" consulting, PE funds, or operating-company CXO roles. Your shosha relationship capital and international ops experience travels well.

Q. Sogo shosha vs specialty trading house — which is safer for job hunting?
A. Safety order: Big Five sogo shosha >> large specialty trading houses > mid-tier specialty. Sogo shosha diversify AI / policy risk via portfolio. Within specialty, choose firms that own brands, products, or specialized technology. Information-brokerage-centric specialty firms are exposed by 2030.

Q. Do shosha-man salaries hold?
A. Seniors: hold or rise; juniors: relative growth slows as hiring shrinks. The Big Five average comp (¥15–18M) holds, but "everyone goes up by tenure" shrinks; promotion concentrates on those with "AI fluency + investment judgment." ¥8M new-grad starting comp holds.

Q. Safest function inside a trading house?
A. Investment planning / corporate development (large M&A, emerging-market investment, geopolitical risk) is the safest. Operations on the downstream side (convenience stores, food, housing, healthcare) has big AI-leverage surface and growth. The riskiest is junior sales-team and clerical work. Aim assignments at "investment / corporate dev / downstream subsidiary secondment" — that's the 2026 right answer.