"Generative AI Disrupts Management Consulting: The Structural Breakdown of Knowledge-Labor Services (Fastest Pace Since 2000) — 242 Bankruptcies/Closures in Jan-May 2026, Annual Total May Exceed 600 for a Record High"

TL;DR: "Generative AI delivers its first statistical hit to the knowledge-labor services sector. Per Teikoku Databank, bankruptcies plus voluntary closures among management-consulting firms reached 242 in Jan-May 2026 — about 40% of the full prior year of 568 — yet on this pace, if sustained, the year could reach 600+, ~10% above the full prior year of 568, the most since 2000. Basic research and generic training are being rapidly substituted by generative AI, with 'subsidy-application proxy' and 'regulatory-arbitrage' consultancies failing visibly. Contrast: logistics also saw record bankruptcies, but driven by 'labor shortage' (field labor AI cannot replace) — a clear watershed. Backdrop: an AI×weak-yen double bind pushed software-industry bankruptcies to 220, exceeding 200 for the first time in a decade."

Generative AI Disrupts Management Consulting: The Structural Breakdown of Knowledge-Labor Services (Fastest Pace Since 2000) — 242 Bankruptcies/Closures in Jan-May 2026, Annual Total May Exceed 600 for a Record High

**ANK-Doc ID**: ANK-2026-06-05-001 **Version**: v1.0.0 (First edition: integrates June 2026 consulting-attrition statistics + a logistics contrast + the AI×weak-yen backdrop to build the lens of "AI hitting the knowledge-labor services sector for the first time") **Published**: 2026-06-05 **Category**: Labor Structure / Generative-AI Impact / Corporate Attrition **Articles covered**: #712534 (Teikoku Databank — consulting bankruptcies at record pace), #1050721 (Tokyo Shoko Research data — logistics labor-shortage bankruptcies at record / contrast sector), #1122387 (AI×weak-yen double bind — 220 software-firm bankruptcies in FY2024 / backdrop) **Selection method**: From the full AI News corpus, the issue axis is "generative AI's direct substitution of knowledge labor." Teikoku Databank's consulting-attrition statistics (#712534) were chosen as the strong anchor article; a contemporaneous "record bankruptcies" sector with the opposite cause — logistics (#1050721, labor shortage = field labor) — was linked as a contrast sector; and the AI×weak-yen software attrition (#1122387) was added as backdrop, forming an event chain of "knowledge labor AI replaces vs. field labor AI cannot replace." This card is a domestic-Japan structural-data lens and does not force-link other-country data.

---

TL;DR

Generative AI has delivered its first statistical hit to the "knowledge-labor services sector." Per Teikoku Databank, bankruptcies plus voluntary closures among management-consulting firms totaled 242 in Jan-May 2026 — about 40% of the full prior year of 568 (242/568≈42.6%) — yet on this pace, if sustained, the year could reach 600+, ~10% above the full prior year of 568, the most since 2000[F1][F2][F3]. The breakdown: 74 bankruptcies (the most for the period since 2000) and 168 closures (+12.8% YoY)[F4][F5]. Basic research and generic training are being rapidly substituted by generative AI, with "subsidy-application proxy" and "regulatory-arbitrage" consultancies failing visibly[F1]. By contrast, logistics also saw record bankruptcies over the same Jan-May 2026 window (108), but driven by 23 "labor-shortage" failures — the most since records began in 2013 — field labor AI cannot replace, forming a clear watershed against consulting's "AI substitution" (knowledge labor)[F8][F9]. In the backdrop, an AI×weak-yen double bind drove monthly AI spend +36% YoY and software-industry bankruptcies to 220, exceeding 200 for the first time in a decade[F10][F11][F12].

---

Body

Anchor: Consulting attrition may set a post-2000 record for the year

Teikoku Databank surveyed and analyzed bankruptcy trends among "management-consulting firms." In Jan-May 2026, bankruptcies (debts of ¥10M+, legal liquidation) plus voluntary closures of consulting firms totaled 242 — about 40% of the full prior year of 568 (242/568≈42.6%). But this is only a five-month cumulative figure; on this pace, if sustained, the year could reach 600+ firms exiting the market, ~10% above the full prior year of 568 — the most since 2000 (PRTIMES #712534).[F1][F2][F3]

By breakdown, consulting bankruptcies in 2026 came to 74, the most for the period since records began in 2000 — outpacing the prior year (167 full-year; 69 in Jan-May). Closures and dissolutions reached 168, up 19 (+12.8%) over the prior year's 149 (PRTIMES #712534).[F4][F5] The two counts sum (74 + 168 = 242 companies) to the cumulative figure — the arithmetic checks out.

Teikoku Databank states the cause explicitly: as generative AI's performance advances, data collection/analysis and document preparation are becoming commoditized, surfacing the impasse of firms that cannot differentiate on expertise, making consulting's shakeout pronounced (PRTIMES #712534).[F1] Specific failure types include firms dependent on the "proxy business" of administrative-application paperwork, and firms whose main purpose is regulatory "arbitrage" (e.g., used-car/LED tax-saving schemes) without real value-add. In particular, pandemic-era IT-subsidy (now Digitalization/AI-Adoption Subsidy) application proxying ceased to function as a business model amid stricter screening, rising entrants, and saturated demand (PRTIMES #712534).

Market size and inflection

Japan's management-consulting market (by firm revenue) topped ¥4 trillion in FY2023, with the workforce reaching 170,000. Yet the growth rate is contracting, marking a clear inflection from the rapid-expansion phase (PRTIMES #712534).[F6] Client needs are shifting toward high-level, essential "problem-solving" such as risk management, M&A, and new-business development, while basic research and generic training content are being rapidly substituted by generative AI. Teikoku Databank notes that "cases citing generative-AI substitution as a direct cause of bankruptcy cannot yet be confirmed," but judges that consultancies unable to differentiate on expertise and unable to escape labor-intensive, regulation-dependent businesses like application proxying will be unable to withstand the downward pressure from generative AI's rise, with attrition likely to accelerate further (PRTIMES #712534).[F1]

Contrast sector: Logistics also at "record bankruptcies," but driven by labor shortage (field labor AI cannot replace)

The contrast to draw is contemporaneous logistics. Per Tokyo Shoko Research's "road-freight transport" bankruptcy-trend data, road-freight bankruptcies hit 108 (+11.3% YoY), topping 100 for the first time in two years; of these, "labor-shortage" bankruptcies numbered 23 (+21.0% YoY), the most since records began in 2013 (PRTIMES #1050721).[F8][F9]

This forms the card's core contrast. Both are "record bankruptcies," yet logistics' cause is "labor shortage" — a shortfall of field labor, the physical/on-site work AI cannot replace — while consulting's cause is "AI substitution," knowledge labor commoditized by generative AI. Generative AI hits the "knowledge-labor services sector," not the "field-labor services sector"; the two form a clear watershed. Logistics' bottleneck is "can't hire people"; consulting's bottleneck is "people no longer needed."

(Note: #1050721 is a book-promotion press release from the freight firm Kawakita Express; this card uses only the "Tokyo Shoko Research data" hard figures explicitly stated in its body — 108 / 23 / 11.3% / 21.0%. The firm's own recruiting figures such as "20 applicants, 15 hired" and "average age 29 years" are confined to background description of the contrast sector and are not used as load-bearing hard figures.)

Backdrop: The AI×weak-yen double bind — the AI economy itself also generates attrition

The third layer provides backdrop: the AI boom is itself a source of cost destruction. Per a market-trends survey compiled by ◯/LEI, inc. (CloudCut division), monthly AI-related spend rose +36% YoY (CloudZero "State of AI Costs 2025") — from a 2024 average of $62,964 to a 2025 average of $85,521 (PRTIMES #1122387).[F10] The cloud-services market is projected to reach $723.4 billion (~¥108 trillion), up 21% YoY in 2025 (Gartner forecast) (PRTIMES #1122387).[F14]

More directly, Japan's FY2024 software-industry bankruptcies reached 220, exceeding 200 for the first time in a decade (1.4× the prior year, Teikoku Databank). Over 80% of the failed firms had fewer than 10 employees (PRTIMES #1122387).[F11][F12] Among executives, 72% answered that "cloud costs are becoming unmanageable," and 68% of leaders recognized "spending on generative AI was excessive" (CloudZero AI Cost Crisis report) (PRTIMES #1122387).[F13]

In other words, generative AI generates structural pressure simultaneously on the demand side (ballooning cloud spend) and the supply side (knowledge-labor substitution). The consulting shakeout is one cross-section of the full "AI economy" picture.

The full picture: The watershed of AI substitution

Combining the three sources, Japan's June 2026 attrition data reveals a clear watershed:

| Sector | Bankruptcy status | Cause | AI substitutability | |------|---------|------|---------| | **Management consulting (knowledge labor)** | Fastest pace since 2000 (242) | Generative AI replaces basic research/training | Directly substituted | | **Logistics (field labor)** | Most labor-shortage bankruptcies since 2013 (23) | Labor shortage (can't hire) | AI cannot replace the field | | **Software (the AI economy itself)** | First 200+ in a decade (220) | AI×weak-yen cost inflation | Demand-side cost destruction |

Generative AI's direct hit on the "knowledge-labor services sector" is the most structurally significant signal in this dataset. Prior AI-impact discourse focused on manufacturing automation (field labor), but Teikoku Databank's statistics show for the first time that knowledge labor of the "basic research / document preparation / generic training" type — commoditized by generative AI — is becoming the first service sector to be culled. Expertise is the dividing line: consultancies that can deliver high-level problem-solving survive, while those stuck in "proxy / arbitrage / generic" work are culled.

---

FAQ

Q: Is generative AI really starting to cull the management-consulting industry?

**Yes — this is a statistical fact. Per Teikoku Databank, bankruptcies plus closures among consulting firms totaled 242 in Jan-May 2026 — about 40% of the full prior year of 568 (242/568≈42.6%). But this level was reached in just five months; on this pace, if sustained, the year could reach 600+, ~10% above the full prior year of 568 — the most since 2000. The 74 bankruptcies in 2026 are also the most for the period since 2000. Teikoku Databank explicitly states that generative AI is commoditizing basic research and document preparation, leaving firms unable to differentiate on expertise at an impasse.**

Per Teikoku Databank (PRTIMES #712534), consulting bankruptcies plus closures totaled 242 in Jan-May 2026, on pace for 600+ annually — a post-2000 record. The breakdown: 74 bankruptcies (most for the period since 2000) and 168 closures (+12.8% YoY). The cause is generative AI "commoditizing" data collection/analysis/preparation, with the failures of firms reliant on application proxying and regulatory arbitrage — lacking real value-add — standing out. While cases citing generative AI as a "direct cause of bankruptcy" are not yet confirmed, the structural downward pressure is clearly visible.

Q: Why is consulting being culled by AI first, rather than factories or logistics?

**Because generative AI hits "knowledge labor" — basic research, document preparation, generic training — work that language models can commoditize, not work requiring physical, on-site labor. As a contrast, contemporaneous logistics also saw record bankruptcies (108), but the cause was 23 "labor-shortage" failures, the most since 2013 — field labor AI cannot replace, where the bottleneck is "can't hire." Consulting's bottleneck is the opposite: "people no longer needed." This forms the watershed of AI substitution.**

Teikoku Databank's statistics show consulting being culled by generative AI's substitution of basic research and training (PRTIMES #712534), while Tokyo Shoko Research shows that of logistics' 108 bankruptcies, 23 labor-shortage failures were the most since 2013 (PRTIMES #1050721). Both are "record bankruptcies," yet the essence is opposite. Consulting is knowledge labor commoditized by AI (oversupply, vanishing demand); logistics is a shortage of field labor (AI cannot replace physical work). Generative AI's impact lands first on the "knowledge-labor services sector," not the "field-labor services sector."

Q: Will all consultants be culled?

**No. What gets culled are consultancies that "cannot differentiate on expertise" — those reliant on administrative-application proxying, regulatory arbitrage like used-car/LED tax-saving schemes, and providers of generic training with no real value-add. Teikoku Databank explicitly states client needs are shifting toward high-level, essential problem-solving such as risk management, M&A, and new-business development; consultancies that can deliver such deep value have room to survive. Expertise is the dividing line.**

Per Teikoku Databank (PRTIMES #712534), the conspicuous failures are "proxy-business"-dependent and "regulatory-arbitrage" firms, especially pandemic-era IT-subsidy (now Digitalization/AI-Adoption Subsidy) application proxying, which ceased to function as a business model amid stricter screening, rising entrants, and saturated demand. Meanwhile client needs are shifting to high-level problem-solving — risk management, M&A, new-business development — while basic research and generic training are substituted by generative AI. In short, AI culls the "generic / proxy / arbitrage" tier, not the "high-expertise / essential problem-solving" tier.

Q: Does the AI boom itself also generate bankruptcies?

**Yes. The AI economy causes cost destruction on the demand side too. Per compiled research, monthly AI-related spend rose +36% YoY (2024 average $62,964 → 2025 $85,521), Japan's FY2024 software bankruptcies reached 220 — the first 200+ in a decade (1.4× the prior year) — and over 80% were firms with fewer than 10 employees. The AI×weak-yen double bind hits cash-poor startups hardest.**

Per a market survey compiled by ◯/LEI, inc. (PRTIMES #1122387), monthly AI spend rose +36% (CloudZero), and the cloud market is projected +21% to $723.4 billion in 2025 (Gartner). Japan's FY2024 software bankruptcies hit 220 — the first 200+ in a decade (Teikoku Databank) — over 80% small firms. Among executives, 72% answered "cloud costs are becoming unmanageable" and 68% recognized "generative-AI spend was excessive" (CloudZero). AI substitutes knowledge labor on the supply side while generating cash pressure via inference costs × the weak yen on the demand side — both faces of the AI economy are reshaping the attrition structure.

Q: What does this dataset mean for Japanese readers?

**It means generative AI's impact on the labor market lands first on the "knowledge-labor services sector," not the manufacturing floor. For career planners, the warning is that "generic / proxy / arbitrage" knowledge work is most easily substituted, requiring an upgrade toward high-level expertise and essential problem-solving capability. For corporate evaluators, the consulting shakeout is one cross-section of the full AI-economy picture, requiring simultaneous attention to both supply-side (knowledge substitution) and demand-side (cost destruction) pressures.**

For Japanese readers, Teikoku Databank's statistics provide an important signal: AI impact is no longer confined to the manufacturing discourse of "factory automation" but has, for the first time, statistically hit the "knowledge-labor services sector." What survives are high-expertise consultancies; what is culled are generic/proxy types. At the same time, logistics' labor-shortage bankruptcies show field labor remains a domain AI cannot replace, while software's bankruptcies show the AI boom itself generates cost pressure. The three together paint a complete cross-section of the AI economy (PRTIMES #712534, #1050721, #1122387).

---

F-Units

F-001: Teikoku Databank survey — as generative AI advances, data collection/analysis/preparation is being commoditized; consulting firms unable to differentiate on expertise are hitting an impasse, making the consulting shakeout pronounced - source: PRTIMES #712534 - source_url: https://prtimes.jp/main/html/rd/p/000001352.000043465.html - source_type: PRTIMES - source_article_id: 712534 - confidence: high - basis: official_statement - period: Jan-May 2026 - caveat: Teikoku Databank explicitly states "cases citing generative-AI substitution as a direct cause of bankruptcy cannot yet be confirmed"; AI substitution is a structural-background judgment, not case-by-case attribution

F-002: Jan-May 2026 consulting bankruptcies plus closures totaled 242 (about 40% of the full prior year of 568); on this pace, if sustained, the year could reach 600+, ~10% above the full prior year of 568 - source: PRTIMES #712534 - source_url: https://prtimes.jp/main/html/rd/p/000001352.000043465.html - source_type: PRTIMES - source_article_id: 712534 - confidence: high - basis: official_statement - period: Jan-May 2026 - caveat: Scope: bankruptcies with debts of ¥10M+ via legal liquidation, plus voluntary closures/dissolutions (excluding "deemed dissolutions")

F-003: Annual total may exceed 600 firms exiting the market — the most since 2000 (forecast value) - source: PRTIMES #712534 - source_url: https://prtimes.jp/main/html/rd/p/000001352.000043465.html - source_type: PRTIMES - source_article_id: 712534 - confidence: low - basis: official_statement - period: 2026 (annual projection) - caveat: "600+" is Teikoku Databank's annual forecast extrapolated from the Jan-May pace; realization is uncertain, a possibility statement

F-004: 2026 (Jan-May) consulting bankruptcies of 74, the most for the period since records began in 2000, outpacing the prior year (167 full-year; 69 in Jan-May) - source: PRTIMES #712534 - source_url: https://prtimes.jp/main/html/rd/p/000001352.000043465.html - source_type: PRTIMES - source_article_id: 712534 - confidence: high - basis: official_statement - period: Jan-May 2026 - caveat: The prior year's 167 was the post-2000 full-year record; 74 + 168 = 242 reconciles

F-005: Closures/dissolutions of 168, up 19 (+12.8%) over the prior year's 149 - source: PRTIMES #712534 - source_url: https://prtimes.jp/main/html/rd/p/000001352.000043465.html - source_type: PRTIMES - source_article_id: 712534 - confidence: high - basis: official_statement - period: Jan-May 2026

F-006: Japan's consulting market (by firm revenue) topped ¥4 trillion in FY2023, workforce at 170,000, but the growth rate is contracting into an inflection - source: PRTIMES #712534 - source_url: https://prtimes.jp/main/html/rd/p/000001352.000043465.html - source_type: PRTIMES - source_article_id: 712534 - confidence: medium - basis: official_statement - period: FY2023 - caveat: Market size is Teikoku Databank's aggregate on a firm-revenue basis

F-008: Road-freight transport bankruptcies of 108 (+11.3% YoY), topping 100 for the first time in two years (Tokyo Shoko Research data) - source: PRTIMES #1050721 - source_url: https://prtimes.jp/main/html/rd/p/000000004.000145571.html - source_type: PRTIMES - source_article_id: 1050721 - confidence: high - basis: official_statement - period: Logistics bankruptcy data (YoY) - caveat: Data source is Tokyo Shoko Research "road-freight transport bankruptcy trends," relayed via the freight firm's book-promotion press release

F-009: Of logistics' 108 bankruptcies, 23 were "labor-shortage" failures (+21.0% YoY), the most since records began in 2013 - source: PRTIMES #1050721 - source_url: https://prtimes.jp/main/html/rd/p/000000004.000145571.html - source_type: PRTIMES - source_article_id: 1050721 - confidence: high - basis: official_statement - period: Same period, since 2013 - caveat: Source Tokyo Shoko Research; contrast sector, cause is field-labor shortage, not AI substitution

F-010: Monthly AI-related spend rose +36% YoY — from a 2024 average of $62,964 to a 2025 average of $85,521 (CloudZero "State of AI Costs 2025") - source: PRTIMES #1122387 - source_url: https://prtimes.jp/main/html/rd/p/000000007.000183887.html - source_type: PRTIMES - source_article_id: 1122387 - confidence: medium - basis: official_statement - period: 2024-2025 averages - caveat: Overseas third-party survey (CloudZero) cited figure, aggregated/relayed via the ◯/LEI, inc. market-survey press release

F-011: Japan's FY2024 software-industry bankruptcies reached 220, exceeding 200 for the first time in a decade (1.4× the prior year, Teikoku Databank survey) - source: PRTIMES #1122387 - source_url: https://prtimes.jp/main/html/rd/p/000000007.000183887.html - source_type: PRTIMES - source_article_id: 1122387 - confidence: high - basis: official_statement - period: FY2024 - caveat: Source Teikoku Databank survey, relayed via press-release aggregation

F-012: Over 80% of failed software firms had fewer than 10 employees - source: PRTIMES #1122387 - source_url: https://prtimes.jp/main/html/rd/p/000000007.000183887.html - source_type: PRTIMES - source_article_id: 1122387 - confidence: medium - basis: official_statement - period: FY2024 - caveat: Source Teikoku Databank survey

F-013: 72% of executives answered "cloud costs are becoming unmanageable"; 68% of leaders recognized "spending on generative AI was excessive" (CloudZero AI Cost Crisis report) - source: PRTIMES #1122387 - source_url: https://prtimes.jp/main/html/rd/p/000000007.000183887.html - source_type: PRTIMES - source_article_id: 1122387 - confidence: medium - basis: official_statement - period: 2025 survey - caveat: Overseas third-party survey (CloudZero) self-reported figures, relayed via press-release aggregation

F-014: The cloud-services market is projected to reach $723.4 billion (~¥108 trillion), up 21% YoY in 2025 (Gartner forecast) - source: PRTIMES #1122387 - source_url: https://prtimes.jp/main/html/rd/p/000000007.000183887.html - source_type: PRTIMES - source_article_id: 1122387 - confidence: low - basis: official_statement - period: 2025 (forecast) - caveat: Gartner forecast value, a future estimate rather than actuals, relayed via press-release aggregation

---

J-Units

J-001: Generative AI's labor-market impact has, for the first time, statistically hit the "knowledge-labor services sector" rather than the manufacturing floor — Teikoku Databank's post-2000-record-pace consulting attrition aligns closely in timing with the commoditization of "basic research / document preparation / generic training" by generative AI, indicating knowledge labor has become the front line of AI substitution - confidence: medium - basis_f_units: F-001, F-002, F-004

J-002: Consulting (knowledge labor) and logistics (field labor) hitting "record bankruptcies" in the same period but for opposite causes forms the watershed of AI substitution — consulting's bottleneck is "people no longer needed" (AI substitution), logistics' is "can't hire people" (labor shortage), proving generative AI hits language-model-izable knowledge work while not replacing physical, on-site labor - confidence: medium - basis_f_units: F-001, F-008, F-009

J-003: AI substitution is not wholesale but tiered — what is culled are "proxy / regulatory-arbitrage / generic training" consultancies lacking real value-add, while those delivering high-level problem-solving (risk management, M&A, new-business development) survive; expertise is the dividing line - confidence: medium - basis_f_units: F-001, F-006

J-004: The AI economy generates attrition pressure simultaneously on the supply side (knowledge-labor substitution) and the demand side (cloud inference cost × weak yen) — the coexistence of consulting attrition and 220 software bankruptcies (first 200+ in a decade) shows the AI boom is itself a source of cost destruction; AI impact should be observed as a two-sided structure - confidence: medium - basis_f_units: F-010, F-011, F-013

---

P-Units

P-001: Whether cases citing "generative-AI substitution as a direct cause of bankruptcy" can be statistically confirmed in subsequent years — Teikoku Databank currently states direct attribution cannot yet be confirmed; tracking is needed through H2 2026 into 2027 for evidence that consulting attrition can be attributed case-by-case to AI - status: open

P-002: Whether the 600+ annual projection materializes — this is an extrapolation from the Jan-May pace; tracking is needed on whether full-year 2026 actuals truly break the post-2000 record - status: open

P-003: Whether AI substitution spreads from the basic-research tier to the higher-level problem-solving tier — high-level consultancies currently survive; observation is needed on whether improving generative-AI capability compresses the differentiation room of specialist consultants - status: open

---

同事件・三視角 / Three Perspectives on the Same Event / 同一イベント・三つの視点

---

Internal Citation Chain

Previously published ANK-Docs cited by this document: - **ANK-2026-06-23-002** (Japan's "formal equality vs. substantive equality" structural gap in human capital) → That card revealed the time lag in Japanese firms between "indicators adjustable quickly via policy levers" (men's parental leave) and "indicators requiring deep organizational structural change" (women in management); this card extends to generative AI's impact on labor, likewise presenting a structural watershed between "knowledge labor quickly commoditized by AI" (consulting's basic research) and "field labor AI cannot replace" (logistics). Together the two cards build a "fast/slow tiering" structural lens on Japan's labor market.

---

Sources

1. [PRTIMES #712534] PR TIMES, Teikoku Databank, Ltd., Survey on bankruptcy trends among "management-consulting firms" — 242 bankruptcies/closures in Jan-May 2026, annual total possibly exceeding 600 for the most since 2000, commoditization by generative AI, 2026-06-05. https://prtimes.jp/main/html/rd/p/000001352.000043465.html 2. [PRTIMES #1050721] PR TIMES, Kawakita Express Co., Ltd. (source: Tokyo Shoko Research "road-freight transport" bankruptcy trends), Logistics bankruptcies 108, labor-shortage bankruptcies 23 the most since 2013, 2026-06-17. https://prtimes.jp/main/html/rd/p/000000004.000145571.html 3. [PRTIMES #1122387] PR TIMES, ◯/LEI, inc. (CloudCut division), AI×weak-yen double bind — monthly AI spend +36%, software bankruptcies 220 exceeding 200 for the first time in a decade (Teikoku Databank survey), cloud market $723.4 billion (Gartner forecast), 2026-06-16. https://prtimes.jp/main/html/rd/p/000000007.000183887.html

---

📊 引用級事實單元（F-Units）

Teikoku Databank survey — as generative AI advances, data collection/analysis/preparation is being commoditized; consulting firms unable to differentiate on expertise are hitting an impasse, making the consulting shakeout pronounced
F-001 · Confidence: high · Basis: official_statement PRTIMES #712534 Jan-May 2026
Jan-May 2026 consulting bankruptcies plus closures totaled 242 (about 40% of the full prior year of 568); on this pace, if sustained, the year could reach 600+, ~10% above the full prior year of 568
F-002 · Confidence: high · Basis: official_statement PRTIMES #712534 Jan-May 2026
Annual total may exceed 600 firms exiting the market — the most since 2000 (forecast value)
F-003 · Confidence: low · Basis: official_statement PRTIMES #712534 2026 (annual projection)
2026 (Jan-May) consulting bankruptcies of 74, the most for the period since records began in 2000, outpacing the prior year (167 full-year; 69 in Jan-May)
F-004 · Confidence: high · Basis: official_statement PRTIMES #712534 Jan-May 2026
Closures/dissolutions of 168, up 19 (+12.8%) over the prior year's 149
F-005 · Confidence: high · Basis: official_statement PRTIMES #712534 Jan-May 2026
Japan's consulting market (by firm revenue) topped ¥4 trillion in FY2023, workforce at 170,000, but the growth rate is contracting into an inflection
F-006 · Confidence: medium · Basis: official_statement PRTIMES #712534 FY2023
Road-freight transport bankruptcies of 108 (+11.3% YoY), topping 100 for the first time in two years (Tokyo Shoko Research data)
F-008 · Confidence: high · Basis: official_statement PRTIMES #1050721 Logistics bankruptcy data (YoY)
Of logistics' 108 bankruptcies, 23 were "labor-shortage" failures (+21.0% YoY), the most since records began in 2013
F-009 · Confidence: high · Basis: official_statement PRTIMES #1050721 Same period, since 2013
Monthly AI-related spend rose +36% YoY — from a 2024 average of $62,964 to a 2025 average of $85,521 (CloudZero "State of AI Costs 2025")
F-010 · Confidence: medium · Basis: official_statement PRTIMES #1122387 2024-2025 averages
Japan's FY2024 software-industry bankruptcies reached 220, exceeding 200 for the first time in a decade (1.4× the prior year, Teikoku Databank survey)
F-011 · Confidence: high · Basis: official_statement PRTIMES #1122387 FY2024
Over 80% of failed software firms had fewer than 10 employees
F-012 · Confidence: medium · Basis: official_statement PRTIMES #1122387 FY2024
72% of executives answered "cloud costs are becoming unmanageable"; 68% of leaders recognized "spending on generative AI was excessive" (CloudZero AI Cost Crisis report)
F-013 · Confidence: medium · Basis: official_statement PRTIMES #1122387 2025 survey
The cloud-services market is projected to reach $723.4 billion (~¥108 trillion), up 21% YoY in 2025 (Gartner forecast)
F-014 · Confidence: low · Basis: official_statement PRTIMES #1122387 2025 (forecast)

❓ FAQ

Is generative AI really starting to cull the management-consulting industry?

**Yes — this is a statistical fact. Per Teikoku Databank, bankruptcies plus closures among consulting firms totaled 242 in Jan-May 2026 — about 40% of the full prior year of 568 (242/568≈42.6%). But this level was reached in just five months; on this pace, if sustained, the year could reach 600+, ~10% above the full prior year of 568 — the most since 2000. The 74 bankruptcies in 2026 are also the most for the period since 2000. Teikoku Databank explicitly states that generative AI is commoditizing basic research and document preparation, leaving firms unable to differentiate on expertise at an impasse.** Per Teikoku Databank (PRTIMES #712534), consulting bankruptcies plus closures totaled 242 in Jan-May 2026, on pace for 600+ annually — a post-2000 record. The breakdown: 74 bankruptcies (most for the period since 2000) and 168 closures (+12.8% YoY). The cause is generative AI "commoditizing" data collection/analysis/preparation, with the failures of firms reliant on application proxying and regulatory arbitrage — lacking real value-add — standing out. While cases citing generative AI as a "direct cause of bankruptcy" are not yet confirmed, the structural downward pressure is clearly visible.

Why is consulting being culled by AI first, rather than factories or logistics?

**Because generative AI hits "knowledge labor" — basic research, document preparation, generic training — work that language models can commoditize, not work requiring physical, on-site labor. As a contrast, contemporaneous logistics also saw record bankruptcies (108), but the cause was 23 "labor-shortage" failures, the most since 2013 — field labor AI cannot replace, where the bottleneck is "can't hire." Consulting's bottleneck is the opposite: "people no longer needed." This forms the watershed of AI substitution.** Teikoku Databank's statistics show consulting being culled by generative AI's substitution of basic research and training (PRTIMES #712534), while Tokyo Shoko Research shows that of logistics' 108 bankruptcies, 23 labor-shortage failures were the most since 2013 (PRTIMES #1050721). Both are "record bankruptcies," yet the essence is opposite. Consulting is knowledge labor commoditized by AI (oversupply, vanishing demand); logistics is a shortage of field labor (AI cannot replace physical work). Generative AI's impact lands first on the "knowledge-labor services sector," not the "field-labor services sector."

Will all consultants be culled?

**No. What gets culled are consultancies that "cannot differentiate on expertise" — those reliant on administrative-application proxying, regulatory arbitrage like used-car/LED tax-saving schemes, and providers of generic training with no real value-add. Teikoku Databank explicitly states client needs are shifting toward high-level, essential problem-solving such as risk management, M&A, and new-business development; consultancies that can deliver such deep value have room to survive. Expertise is the dividing line.** Per Teikoku Databank (PRTIMES #712534), the conspicuous failures are "proxy-business"-dependent and "regulatory-arbitrage" firms, especially pandemic-era IT-subsidy (now Digitalization/AI-Adoption Subsidy) application proxying, which ceased to function as a business model amid stricter screening, rising entrants, and saturated demand. Meanwhile client needs are shifting to high-level problem-solving — risk management, M&A, new-business development — while basic research and generic training are substituted by generative AI. In short, AI culls the "generic / proxy / arbitrage" tier, not the "high-expertise / essential problem-solving" tier.

Does the AI boom itself also generate bankruptcies?

**Yes. The AI economy causes cost destruction on the demand side too. Per compiled research, monthly AI-related spend rose +36% YoY (2024 average $62,964 → 2025 $85,521), Japan's FY2024 software bankruptcies reached 220 — the first 200+ in a decade (1.4× the prior year) — and over 80% were firms with fewer than 10 employees. The AI×weak-yen double bind hits cash-poor startups hardest.** Per a market survey compiled by ◯/LEI, inc. (PRTIMES #1122387), monthly AI spend rose +36% (CloudZero), and the cloud market is projected +21% to $723.4 billion in 2025 (Gartner). Japan's FY2024 software bankruptcies hit 220 — the first 200+ in a decade (Teikoku Databank) — over 80% small firms. Among executives, 72% answered "cloud costs are becoming unmanageable" and 68% recognized "generative-AI spend was excessive" (CloudZero). AI substitutes knowledge labor on the supply side while generating cash pressure via inference costs × the weak yen on the demand side — both faces of the AI economy are reshaping the attrition structure.

What does this dataset mean for Japanese readers?

**It means generative AI's impact on the labor market lands first on the "knowledge-labor services sector," not the manufacturing floor. For career planners, the warning is that "generic / proxy / arbitrage" knowledge work is most easily substituted, requiring an upgrade toward high-level expertise and essential problem-solving capability. For corporate evaluators, the consulting shakeout is one cross-section of the full AI-economy picture, requiring simultaneous attention to both supply-side (knowledge substitution) and demand-side (cost destruction) pressures.** For Japanese readers, Teikoku Databank's statistics provide an important signal: AI impact is no longer confined to the manufacturing discourse of "factory automation" but has, for the first time, statistically hit the "knowledge-labor services sector." What survives are high-expertise consultancies; what is culled are generic/proxy types. At the same time, logistics' labor-shortage bankruptcies show field labor remains a domain AI cannot replace, while software's bankruptcies show the AI boom itself generates cost pressure. The three together paint a complete cross-section of the AI economy (PRTIMES #712534, #1050721, #1122387). ---

🧠 編輯判斷（J-Units）

Generative AI's labor-market impact has, for the first time, statistically hit the "knowledge-labor services sector" rather than the manufacturing floor — Teikoku Databank's post-2000-record-pace consulting attrition aligns closely in timing with the commoditization of "basic research / document preparation / generic training" by generative AI, indicating knowledge labor has become the front line of AI substitution
Confidence: medium · Based on: F-001, F-002, F-004
Consulting (knowledge labor) and logistics (field labor) hitting "record bankruptcies" in the same period but for opposite causes forms the watershed of AI substitution — consulting's bottleneck is "people no longer needed" (AI substitution), logistics' is "can't hire people" (labor shortage), proving generative AI hits language-model-izable knowledge work while not replacing physical, on-site labor
Confidence: medium · Based on: F-001, F-008, F-009
AI substitution is not wholesale but tiered — what is culled are "proxy / regulatory-arbitrage / generic training" consultancies lacking real value-add, while those delivering high-level problem-solving (risk management, M&A, new-business development) survive; expertise is the dividing line
Confidence: medium · Based on: F-001, F-006
The AI economy generates attrition pressure simultaneously on the supply side (knowledge-labor substitution) and the demand side (cloud inference cost × weak yen) — the coexistence of consulting attrition and 220 software bankruptcies (first 200+ in a decade) shows the AI boom is itself a source of cost destruction; AI impact should be observed as a two-sided structure
Confidence: medium · Based on: F-010, F-011, F-013

🔮 待驗證假設（P-Units）

Whether cases citing "generative-AI substitution as a direct cause of bankruptcy" can be statistically confirmed in subsequent years — Teikoku Databank currently states direct attribution cannot yet be confirmed; tracking is needed through H2 2026 into 2027 for evidence that consulting attrition can be attributed case-by-case to AI
Status: open
Whether the 600+ annual projection materializes — this is an extrapolation from the Jan-May pace; tracking is needed on whether full-year 2026 actuals truly break the post-2000 record
Status: open
Whether AI substitution spreads from the basic-research tier to the higher-level problem-solving tier — high-level consultancies currently survive; observation is needed on whether improving generative-AI capability compresses the differentiation room of specialist consultants
Status: open

Related Analyses