The Decentralisation Paradox: What a Blockchain Collapse Teaches Us About Governing Differently

In 2018, IBM and Maersk launched TradeLens: a blockchain platform meant to digitise global shipping, replacing the paper trail that follows a container from a Vietnamese factory to a Rotterdam warehouse with a single, shared, tamper-proof ledger. It had the backing of two of the largest logistics operators on earth. By 2023, it was dead.

The postmortems mostly blamed adoption: too few competitors were willing to run their supply chain data through infrastructure IBM and Maersk controlled. That is true, but it is not the interesting failure. The interesting failure is why competitors didn’t trust it, and a recent study gives a precise answer. TradeLens was never as decentralised as the word “blockchain” implied, and the gap between the promise and the practice is exactly what killed it.

This matters beyond shipping. It is the same story this series has already told once, about a different technology, and it is worth naming as a pattern rather than leaving as two disconnected anecdotes.

The Paradox: Centralisation in Disguise

Researcher Issam Najati examined TradeLens through the lens of Elinor Ostrom’s design principles for commons governance, the same framework behind the argument in Nature Is Not a Cash Flow and The Cooperatives of the Soil. Ten expert interviews and a documentary review turned up five interrelated governance failures: excessive centralisation of authority, limited procedural transparency and monitoring, rigid rules, minimal small-business involvement, and poor communication between the platform and its supposed community of users.

Najati’s paper builds on a description of the underlying pattern from earlier literature, and the phrase is worth quoting directly:

“The paradox of decentralization: centralization in disguise.” (Sun, 2021, cited in Najati, 2026)

Governance power, the study found, concentrated in the hands of “founding teams, lead developers, protocol maintainers, or major validators” regardless of the technology’s theoretical openness. Ostrom would have recognised this immediately. It is exactly the failure her design principles exist to prevent: some participants exert disproportionate control over a shared resource, and the formal architecture underneath them, however distributed it looks on a whiteboard, doesn’t stop it.

The paper’s most useful line applies well beyond shipping platforms:

“Technological decentralization alone does not substitute for the social and institutional infrastructure required to sustain consensus and collective trust.”

Small freight forwarders, the businesses TradeLens most needed to bring on board to make the network valuable, were structurally excluded from decisions about how the platform worked. Sanctions for rule violations were rigid and automated, with no path for dialogue or proportionate response. There was no accessible mechanism for resolving disputes short of leaving the platform entirely. Every one of these is a direct violation of one of Ostrom’s eight design principles for durable commons: collective-choice arrangements, graduated sanctions, low-cost conflict resolution. TradeLens had a blockchain. It didn’t have commons governance.

The Same Story, With a Token

Regular readers of this site will recognise this shape. Cooperative by Design told a near-identical story about Helium: hundreds of thousands of node operators worldwide, running genuinely distributed wireless infrastructure, in a network that moved packets exactly as advertised, with a corporation underneath it the entire time. Insiders captured the token’s value. Operators carried the capital risk of the hardware. The architecture was decentralised. The governance was not.

That story has kept moving since, and the next chapter hasn’t been announced yet, only planned. The venue is Discord, where a so-called “advisory committee”, the latest costume “community driven” governance has worn at Helium, is actively planning a full shutdown of the IoT network. The committee is drawn from what remains of the largest token-holders and token speculators, which says most of what needs saying about who “community” means at this stage of the network’s life. At its peak, that network counted more than 900,000 registered LoRaWAN gateways, the largest crowd-funded long-range radio network the world has built. Most of that hardware will not be repurposed. A majority of those gateways were already unplugged and turned into e-waste long before any shutdown notice, the token’s decline having done that work years ago. For the minority still running, mostly units built around a Raspberry Pi or with SSH access to their own hardware, there is a real path off Helium: point the gateway at your own ChirpStack backend and keep it running on a network you actually control, rather than one that was renting you the illusion of ownership. A handful of operators are already doing exactly that. It rescues individual gateways, not the network.

The governance picture has not improved either. What remains of the network’s so-called voting power now sits with a handful of insiders and Nova Labs itself, dictating direction as well as holding the votes that were meant to check it, concentration dressed up as process rather than any correction of it. What survives under the Helium name is the “Mobile” network, and it barely resembles the original pitch: largely US-based Wi-Fi hotspots doing carrier data offload, with every carrier deal negotiated by Nova in secret. There will be no worldwide LoRaWAN network left.

None of this is a surprise. Helium’s demise has been predictable for long enough that, when it finally lands, it will read less as news than as a closing argument. That predictability is part of why I have been working out loud on something of my own, SeventyOne. It is not a finished or fully actionable proposal, more a working set of notes trying to take the governance failures this piece keeps returning to, capture by insiders, voting power without accountability, community framing that means the opposite of community, and design them out from the start rather than patch them in after the fact. I am not doing that thinking alone. Some of the people working through it with me are builders I met along Helium’s own decline, people already shipping working pieces of that picture rather than waiting on a finished whitepaper to build from. Governance, on paper at least, is vested in participants through reputation earned by contribution, not token holdings, which is the direct structural answer to a “voting power” that a handful of insiders can otherwise just accumulate and keep.

TradeLens and Helium sit on opposite ends of the blockchain spectrum: one enterprise consortium infrastructure, the other a consumer DePIN network with a speculative token. They failed for the same reason. Decentralised architecture answers the question “where does the data live.” It says nothing about the question “who decides the rules, and who benefits when they change.” Those are governance questions, and blockchain, as a piece of technology, is agnostic about how they get answered. Left unaddressed, they get answered the way they always have: by whoever already holds power over the system’s foundations.

It is worth being fair to the technology here, because the counter-argument matters. A 2021 review by Rozas, Tenorio-Fornés, Díaz-Molina and Hassan, examining blockchain against Ostrom’s principles directly, found real strengths: tokenisation can make boundaries and previously invisible labour visible, smart contracts can automate graduated sanctions, and transparent ledgers genuinely strengthen monitoring. The failure isn’t blockchain as such. It’s specific governance design choices, repeated across projects that had every technical capability to do better and chose concentrated control instead. The same review flags a structural limit worth remembering regardless of which chain or protocol is under discussion: “DAOs are constrained to the digital world.” No smart contract enforces a container actually being loaded onto the right ship.

A Structurally Different Bet

This is the point in the argument where the working notes mentioned above, SeventyOne, stop being coy about their architecture. Part of the bet under discussion, still untested at any real scale, is on a technology that tries to answer the governance question differently rather than better: Holochain.

The distinction matters, because it is easy to hear “Holochain” and assume it means “faster blockchain” or “greener blockchain.” It is neither. Blockchain’s entire model depends on every participant agreeing on one shared, ordered history: a global ledger, validated by consensus, that everyone defers to. Holochain has no equivalent. Each participant, called an agent, keeps their own cryptographically signed local history, called a source chain. When two agents interact, each validates the interaction against rules the specific application defines, then publishes it to a shared distributed hash table for anyone else to independently verify. There is no global ledger to capture, because there is no global ledger. As Holochain’s own documentation puts it: “compute and data live at the edges, which means there’s no infrastructure for you to maintain and defend.”

Governance follows the same logic. Moss, the most developed groupware application built on Holochain, describes itself around self-governing groups, where each group runs its own private peer-to-peer network and chooses its own rules, rather than inheriting governance from however a foundation originally distributed a token. There is no gas fee and no single shared contract whose one bug exposes everyone, because there is no shared chain to attack in the first place. Whether that trade-off is actually better depends entirely on whether a specific group governs itself well, which is Ostrom’s whole point: the technology creates room for good governance. It cannot manufacture it.

The detail that connects this back to the rest of the series is hREA, an implementation of the Valueflows vocabulary running on Holochain’s architecture. The Cooperatives of the Soil already introduced Valueflows as the open standard behind GrowGood, and The Monasteries of the Höllental gestured at open value accounting as one of the hub’s live experiments. hREA is what makes that concrete: the same economic language, running peer-to-peer, with no platform in the middle to capture it.

There is a further question sitting underneath hREA that the wider Valueflows community has not settled yet, and it loops this whole series back to where it started. Valueflows, like the REA (Resource-Event-Agent) accounting model it extends, records economic events that connect two kinds of thing: an agent who acts, and a resource that gets used, transferred, or consumed. That works cleanly for grain, tractors, and contractors. It works less cleanly for a watershed, an aquifer, or a soil profile, the kind of thing Nature Is Not a Cash Flow argues should never be forced into the shape of a tradeable asset in the first place. A river is not an agent in the ordinary sense: it does not negotiate or bear responsibility for what happens to it. But treating it purely as a resource means that once it is abstracted, sold, or metered, the river itself vanishes from the record, and whatever happens to it afterwards becomes somebody else’s problem to document separately. There is early, unsettled thinking in exactly the open-accounting circles building hREA about whether a third category belongs alongside agent and resource: something that can be observed and stewarded without being owned. It is not resolved, and it is not part of the core standard. My own view is that it doesn’t need a new category. Valueflows can already treat a river as an ecological agent, flagged as something nobody can own, that never commits to anything itself and appears in the record only through people accountable for acting on its behalf. That keeps the river visible without forking the standard, and puts the real work where it belongs: in the governance rules about who gets to speak for it. But it is the same question this series keeps circling from different directions: how do you make something visible and governable without first making it property? It is a good sign that the people building the infrastructure are asking it too.

The Honest Caveat

None of this is a solved problem, and pretending otherwise would be the tech utopianism this site tries hard to avoid. Holochain has a well-documented history, stretching from 2018 to 2024, of slow delivery and fair “vaporware” criticism. Moss is still in beta, desktop-only, with a developer ecosystem a fraction the size of Ethereum’s. The honest version of this article is not “Holochain solves what blockchain got wrong.” It is: Holochain removes one specific failure mode, the global ledger that concentrated power can capture, and replaces it with a different, still-early set of trade-offs that haven’t been tested at TradeLens’s scale or under TradeLens’s commercial pressure.

What TradeLens and Helium actually prove is narrower and more useful than “blockchain doesn’t work.” They prove that the question “is this decentralised” is the wrong question to be asking about any shared digital infrastructure, commons-governed or otherwise. The right question is Ostrom’s: who gets to make the rules, who monitors compliance, who has a genuine path to raise a grievance, and what happens when someone breaks the agreement. A network can score badly on every one of those while running flawlessly as a piece of distributed software. Architecture is not governance. It never was.

Sources

• Najati, I. (2026). Governance Failures in Blockchain Ecosystems: Insights from a Collaborative Platform Collapse. Data Science and Management, 9(3), 100193. DOI: 10.1016/j.dsm.2026.100193. The TradeLens case study this post draws on throughout.
• Sun (2021), “Is Blockchain Becoming More Centralized? Evidence on Collusion in the Ethereum Blockchain”, SSRN working paper, cited in Najati (2026). The “paradox of decentralization: centralization in disguise” framing, originating in the literature Najati’s paper reviews rather than Najati’s own coinage.
• Rozas, D., Tenorio-Fornés, A., Díaz-Molina, S., & Hassan, S. (2021). When Ostrom Meets Blockchain: Exploring the Potentials of Blockchain for Commons Governance. SAGE Open, 11(1). DOI: 10.1177/21582440211002526. The fairer, more balanced counterpoint on where blockchain genuinely supports Ostrom’s design principles.
• Ostrom, E. (1990). Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge University Press. The eight design principles applied throughout this piece.
• Holochain (2026). holochain.org. Agent-centric architecture, self-description.
• Lightningrod Labs (2026). Moss. moss.social. Groupware built on Holochain, using the Weave framework for composable group tools.
• hREA Project (2026). hrea.io. Valueflows/REA implementation on Holochain’s agent-centric DHT.
• Gaggl, L. (2026). SeventyOne Network working documents. codeberg.org/leogaggl/seventyone-docs. Working-out-loud notes, not a finished proposal, toward a demand-driven, non-speculative DePIN model built on ValueFlows accounting and reputation-based governance, developed with builders met during Helium’s decline who are already delivering pieces of the same picture.

• Blockchain Governance
• Holochain
• Commons Governance
• Decentralisation
• Ostrom
• DAO
• Data Sovereignty
• Digital Commons