Decentralisation is a powerful design goal and a slippery marketing word. Many systems that use the word still concentrate upgrade keys, parameter control, treasury power, or informal leadership. That is not automatically fraud. It is a governance fact. Investors who skip that fact buy a narrative and inherit a power structure.
Ask who can change the machine
Under stress, who can pause functions, alter collateral factors, redirect incentives, or organise a rescue that socialises losses? If the answer is foggy, your risk model is incomplete. Fog is common. Size should respect fog.
Governance tokens and voting rituals can look like distribution of power while voting power remains concentrated. Read participation, not slogans. Read emergency powers, not only whitepaper ideals.
Traditional analogies help
Corporate governance failures did not wait for blockchains. Controlling shareholders, opaque related-party deals, and board capture are old stories. Use those instincts. Then add crypto-native details: admin keys, multisig membership, upgradeability, oracle control, and venue dependencies.
Regulation intersects governance
Regulators often care about who is accountable when customers are harmed. Builders who treat regulation only as an enemy miss the design constraint: unclear accountability becomes a product defect in institutional eyes. Clearer governance can be a feature, not a surrender.
Education without mythology
Teach students to separate architecture claims from control maps. Celebrate genuine distribution of power when evidence supports it. Refuse to treat the word “decentralised” as a risk rating. Terra-era and other episodes are teaching material when told structurally, without cartoon villain simplification alone.
I can respect experimentation and still demand plain answers about control. Those positions are not contradictory.
Portfolio implication
Due diligence questions that travel well
Who holds upgrade authority? What is the quorum? What emergency actions exist? How are treasuries spent? What conflicts exist between core teams and token holders? Which communications channels are official? You will not always get perfect answers. The quality of answers is itself a signal.
On-chain transparency helps and can mislead. Visible wallets are not the same as understood incentives. Combine on-chain reading with off-chain institutional reading: legal entities, disclosures, and jurisdiction.
When narratives outrun controls
Rapid community growth often precedes governance maturity. Treat that sequence as risk, not as romance. Size smaller while controls catch up—or wait. Waiting is allowed.
For builders, closing governance gaps is product work. Clear accountability can widen the set of institutions willing to engage. That is strategic, not merely defensive.
Believe distributions of power when evidence shows them. Until then, assume concentration and invest as if your assumption might be right.
If governance is opaque, cap size, demand monitoring, and avoid cash-like treatment. If governance is clearer, you still need market and smart-contract risk analysis. Governance clarity is necessary, not sufficient.
Allocators’ minimum governance packet
Before size leaves “tiny,” demand a governance packet: control map, emergency powers, treasury policy, disclosure cadence, and incident history. Incomplete packets cap size automatically. This is not hostility to builders. It is how serious capital speaks.
Narratives scale fast. Governance packets should gate that scale in your book.
Stress reveals the org chart
In calm times, forums and rituals can look participatory. In stress, watch who speaks with authority and whose signatures move parameters. That observed org chart is the one your risk model needed. Update files after stress elsewhere even if you were not involved; learning can be imported.
For educators, teach students to draw control maps from public materials. The exercise builds permanent scepticism toward floating adjectives like “community-run” without evidence.
Power concentrates until proven otherwise. Invest as if that prior might be true.
Stress reveals the org chart
In calm times, forums and rituals can look participatory. In stress, watch who speaks with authority and whose signatures move parameters. That observed org chart is the one your risk model needed. Update files after stress elsewhere even if you were not involved; learning can be imported.
For educators, teach students to draw control maps from public materials. The exercise builds permanent scepticism toward floating adjectives like “community-run” without evidence.
Power concentrates until proven otherwise. Invest as if that prior might be true.
Working notes I keep returning to
On Governance Gaps in Decentralised Narratives, I keep a short list of non-negotiables: write the claim, name the failure, size for survival, and review on a cadence that does not depend on mood. The list is not clever. It is how personal opinion stays accountable to capital.
Markets will keep inventing new costumes for old risks. Process language travels better than costume language. That is why I write these notes in public English for anyone building discipline alongside ambition—including students and builders in Indonesia’s growing fintech conversation.
If you change only one behaviour after reading, make it precommitment: decide the rule before the tape demands an exception. Exceptions can still exist. They should expire in writing.
Narratives decentralise easily. Power does not. Invest accordingly.