This series shares brief case studies of organizations relevant for designing economic mechanisms around ownership of decentralized infrastructure. I intend these posts to be informative, rather than copy-paste recommendations, so I present advantages, disadvantages, and speculations around each model that I learned through my recent research.
Written by Kei Kreutler.
During the mania of 2021, when new DAOs seemed to launch with fanfare every day, a few people started talking about the keiretsu model as a way to manage shared infrastructure.
Keiretsus are "clusters of independently managed firms" defined by shared governance mechanisms, partial cross-shareholding, and interlocking directorates, plus some trade in products.[1]
Keiretsus emerged in post-war Japan upon the forced dissolution of the zaibatsu, which were vertically integrated business conglomerates owned by families. Though historians debate the extent to which zaibatsus continued in other forms, both "horizontal" and "vertical" keiretsus formed in their wake. Often centered around a bank, the member firms of horizontal keiretsus had greater overlap in governance mechanisms, cross-shareholding, and interlocking directorates, whereas vertical keiretsus usually involved one large manufacturing company with suppliers as subsidiaries.[2]
This brief case study includes a high-level overview of horizontal keiretsus that largely draws from Michael L. Gerlach's book Alliance Capitalism.
The Sumitomo Group
The Sumitomo Group illustrates how a horizontal keiretsu operates through its enduring presence in Japan. From its supposed beginnings as a seventeenth-century bookshop founded by a former monk, it grew into a diverse industrial network. While it started with copper refining, the group expanded across industries alongside developing significant financial services.
Three interconnected mechanisms shaped how the Sumitomo Group functioned. The first mechanism revolved around Sumitomo Trust and Bank, which provided financing to member firms as well as acted as a central node in their cross-shareholding financial structure. Popular accounts often overstate the extent of cross-shareholding among member firms within keiretsu, but the reality was more nuanced. Only about 18% of member firms held shares in each other, and of these, the non-financial firms rarely ranked among the top twenty shareholders in each other. Usually only the financial firms, like banks, trusts, and insurance companies within the keiretsu, held significant shares in any given member firm. These financial firms acted as network "'bridges,' connecting member firms through their control over capital and intermediate product markets," much like the Sumitomo Trust and Bank does in the graphic below.[3]
The second key mechanism was their governance approach, centered on regular “presidents’ councils” that gained formal structure in the post-war period. Often named for their meeting schedule, like the Third Wednesday Club, these councils documented proceedings, though records stayed private. They addressed shared concerns around risk mitigation while symbolically strengthening the status-laden group. The Sumitomo Group stood out for its particularly decentralized approach. By the late 1980s, leadership rotated among all twenty member company presidents. Even the Sumitomo Bank president would deliberately take a corner seat, letting leadership emerge naturally based on seniority when needed.[4]
The third key mechanism was extensive resource sharing across the group. They regularly transferred executives between companies in full-time roles, and in some cases, inter-firm research councils formed around emerging technologies. There was also significant trading in intermediate products; between 1982 and 1984, the Sumitomo Group’s trading company participated in transactions with two-thirds of member firms.[5] However, this internal focus sometimes created tension, particularly when foreign joint venture partners preferred to source materials from the open market.
Keiretsus' Relevance for Today
The keiretsu model has evolved significantly since the late 1980s, when Michael L. Gerlach wrote Alliance Capitalism. Their story offers several lessons for internet native organizations.
Japan’s 2021 Corporate Governance Code challenged traditional keiretsu practices by requiring firms to justify their cross-shareholdings. This highlighted a key tension: while cross-shareholding created stable business networks, it also led to capital inefficiencies that some link to Japan’s economic challenges over the past three decades.
The change in capital markets has also fundamentally reshaped how firms access funding. The traditional keiretsu model centered around bank financing, long-term debt relationships, and stable cross-shareholding. J. Mark Ramseyer argues, only partially convincingly, that keiretsu financial firms played a role similar to modern venture capital, funding risky technological innovation while providing emergency liquidity during downturns.[6] In any case, by contrast, modern firms typically have access to multiple funding sources through equity and bonds, use more flexible financial instruments, and form industry-led networks rather than bank-centered ones.
Despite these changes, the keiretsu model’s core insight, that network structures can build resilient business relationships, remains powerful. The Mitsubishi Group's continued global manufacturing success demonstrates how these organizational principles can adapt to modern contexts.[7]
Learnings and Examples
While the keiretsu model may not directly translate to managing shared infrastructure in permissionless markets, certain characteristics of decentralized infrastructure, particularly onchain ownership, could help capture the benefits of keiretsu-style organization while mitigating its traditional disadvantages.
Ultimately, with keiretsus, we can see both the advantages and the certain pitfalls of nascent “squad wealth.”
The model’s approach to managing high-risk investments through networked relationships could offer valuable insights for bootstrapping interdependent projects on layer 2 networks.
Example Scenario
Consider an example L3 network called Keiretsu. This network would operate as a decentralized investment hub, using its treasury to invest in related infrastructure projects in exchange for their tokens.
Similar to how traditional keiretsus used their banking relationships to create stable business networks, Keiretsu would create further network effects through its sequencer membership. Sequencers who demonstrate substantial investment in the network’s infrastructure, specifically, holding a substantial value in infrastructure project tokens that Keiretsu has also invested in, would receive priority for their submissions.
``` // SPDX-License-Identifier: UNLICENSED pragma solidity 0.8.25; import {IsAllowed} from "src/interfaces/IsAllowed.sol"; interface IERC20 { function balanceOf(address account) external view returns (uint256); } interface IOracle { function getETHValue(address token, uint256 amount) external view returns (uint256); } contract KeiretsuSequencing is IsAllowed { address public immutable treasury; address public immutable oracle; address[] public treasuryTokens; uint256 public constant REQUIRED_ETH_VALUE = 32 ether; uint256 public constant MIN_OVERLAP_PERCENT = 33; constructor(address _treasury, address _oracle) { treasury = _treasury; oracle = _oracle; } function isAllowed(address proposer) external view override returns (bool) { // Calculate minimum required overlap (3 or 33%, whichever is lesser) uint256 requiredOverlap = (treasuryTokens.length * MIN_OVERLAP_PERCENT) / 100; requiredOverlap = requiredOverlap < 3 ? requiredOverlap : 3; uint256 overlapCount; uint256 totalValue; // Check each token held by treasury for (uint256 i = 0; i < treasuryTokens.length; i++) { address token = treasuryTokens[i]; uint256 balance = IERC20(token).balanceOf(proposer); uint256 ethValue = IOracle(oracle).getETHValue(token, balance); // Count token if it contributes meaningful value if (ethValue > 0.1 ether) { // Minimum meaningful holding overlapCount++; totalValue += ethValue; } } // Must hold enough overlapping tokens AND their value must be >= 32 ETH return overlapCount >= requiredOverlap && totalValue >= REQUIRED_ETH_VALUE; } } ```
This mechanism transforms sequencers from mere service providers into invested network participants. The Keiretsu L3 could function like a decentrally managed commercial bank, creating bridges between projects, incentivizing cross-shareholding patterns among sequencers, and importantly, maintaining transparent financial relationships onchain.
Ultimately, this approach to relationship building could strengthen the Keiretsu ecosystem’s resilience. In a context where market capitalization often disconnects from productive value, these incentivized interdependencies could provide more stable foundations for long-term development.
[1] Grabowiecki, Jerzy. "Keiretsu Groups," Abstract.
[2] Again, it's worth noting some historians draw distinctions between not only two types of keiretsus but in certain cases, up to five. Additionally, public data on shareholding for the immediate post-war period is limited, which makes it difficult to fully verify these distinctions. Ramseyer, J. Mark. "Cross-Shareholding in the Japanese Keiretsu," 2.
[3] Gerlach, Alliance Capitalism, 109.
[4] Gerlach, Alliance Capitalism, 107-110.
[5] Gerlach, Alliance Capitalism, 143.
[6] Ramseyer, J. Mark. "Cross-Shareholding in the Japanese Keiretsu," 5.
[7] Mitsubishi is the World's Largest Industrial Conglomerate, Bismarck Brief, January 8, 2025.
Sources
Gerlach, Michael L. Alliance Capitalism: The Social Organization of Japanese Business. Berkeley: University of California Press, 1992.
Grabowiecki, Jerzy. "Keiretsu Groups: Their Role in the Japanese Economy and a Reference Point (or Paradigm) for Other Countries." 2006. 1-85.
Ramseyer, J. Mark. "Cross-Shareholding in the Japanese Keiretsu." SSRN Scholarly Paper. Rochester, NY, October 1, 1998. https://papers.ssrn.com/abstract=150552.