The Stability Fallacy in the RIR Argument

The Stability Fallacy in the RIR Argument

Regional Internet Registries like to speak the language of stability. It is a useful word. It sounds technical. It sounds neutral. It sounds above politics, above commerce, and above the dispute itself.

But the word hides the real question.

Stability of what?

AFRINIC and the wider RIR camp often say that my legal challenge threatens registry stability. That claim is not entirely irrational. Litigation against a registry can create uncertainty. Frozen accounts, court orders, disputed elections, receivership, and institutional paralysis are not signs of a healthy registry. Nobody should pretend otherwise.

But this is only half of the ledger. The other half is larger.

Cloud Innovation, LARUS, and my group of companies represent more than 10 million IP addresses. Behind those resources sit millions of websites, hundreds of millions of live end users, and real infrastructure. Hospitals. Airports. Enterprises. Hosting providers. Security systems. Payment systems. Critical public and private services.

If AFRINIC, or any RIR, attempts to revoke, pollute, freeze, or politically contaminate these resources, the stability risk is not theoretical. It is not a matter of minutes in a policy meeting. It is not a philosophical debate about “community.” It is a direct operational threat to running networks.

So the RIR argument is not false because stability does not matter. It is false because it defines stability as the comfort of the registry, not the continuity of the Internet.

There are two types of stability in this dispute.

The first is registry institutional stability. This is the stability of the corporation, its bank account, its board, its office, its elections, its staff, its legal defence, and its self-description as the legitimate regional authority.

The second is routed network and customer-continuity stability. This is the stability of packets moving, services staying online, customers remaining reachable, contracts being honoured, and operators being able to rely on the identifiers around which their infrastructure was built.

AFRINIC speaks mostly about the first. I speak mostly about the second.

That is the difference.

The economic error is simple. RIRs treat the registry as if it were the asset. It is not. The registry is a coordination service. The asset is the live economic use of number resources by operators and customers. The value is not created by a database entry sitting peacefully inside a registry office. The value is created by networks that deploy addresses, customers that rely on them, and markets that build services on top of them.

A registry record describes reality. It does not create it.

Once IPv4 became scarce, financeable, leased, bought, routed, insured, litigated, and embedded into customer infrastructure, the old registry model changed character. What had once been a low-value administrative function became a choke point over valuable operational assets. The RIRs did not update their mandate, liability, representation model, or failover architecture to match that change. They kept the old language of stewardship while exercising new forms of power.

That is the source of instability.

Not the operator who says the emperor has no mandate.
Not the court that asks whether private registry action has legal limits.
Not the customer who wants continuity.
Not the market that treats IPv4 as capital.

The instability comes from a private registry claiming the power to destroy live network assets while bearing little or no proportional liability for the consequences.

In economics this is a textbook hold-up problem. One party makes investments that are specific, sunk, and hard to move. Another party controls a bottleneck after the investment has been made. The bottleneck owner then claims discretionary authority over the asset, while the investor bears the downside. In ordinary markets, this is called a structural risk. In the RIR world, it is called “stewardship.”

The word does not change the structure.

A hospital does not care whether an outage was caused by a noble community process. An airport does not care whether a failed route object was justified by an internal policy interpretation. A website owner does not care whether a registry believes it has defended “the region.” The end user experiences only one thing: the service works or it does not.

That is why the stability argument must be reversed.

AFRINIC says my challenge threatens the stability of the registry. But AFRINIC’s attempted assertion of power over operationally embedded resources threatens the stability of the actual Internet. It converts a registry dispute into a live-infrastructure risk. It creates uncertainty around assets that customers, carriers, enterprises, and public services rely on every day.

This is not stability. It is hostage stability.

The message is: do not challenge the registry, because the registry may collapse. But if the registry can collapse because one member uses courts, then the registry was never stable enough to hold unilateral power over live infrastructure. If a system cannot survive legal scrutiny, asset claims, member dispute, banking pressure, election failure, or ordinary court supervision, the answer is not to give it more immunity. The answer is to reduce its power.

A registry is not a state.
A room is not a mandate.
A policy meeting is not a legislature.
A service region is not a people.
A database contact is not a corporate power of attorney.

And a registry’s desire to preserve its institutional authority is not the same thing as Internet stability.

The RIRs confuse their own survival with the survival of the Internet because their model depends on that confusion. If the registry and the Internet are treated as the same thing, then every challenge to the registry becomes a threat to the Internet. Every lawsuit becomes sabotage. Every demand for liability becomes an attack. Every request for portability becomes capture. Every market actor becomes a danger.

This is institutional self-preservation dressed as engineering.

The real engineering question is narrower. Does a particular action protect uniqueness? Does it prevent duplicate assignment? Does it correct fraud? Does it preserve routing-adjacent accuracy? Does it improve security? Does it maintain interoperability?

If the answer is yes, the registry may have a technical role.

If the answer is no, the registry is not protecting the Internet. It is expanding power.

Commercial use is not a uniqueness problem. Leasing is not a uniqueness problem. Customer geography is not a uniqueness problem. Asset financing is not a uniqueness problem. A member’s business model is not a global invariant. These are matters for operators, customers, contracts, courts, and markets. They are not a licence for a registry to threaten already-running resources.

The RIR system survived for years because the assets were misunderstood. IP addresses were treated as administrative tokens, not capital assets. That fiction is over. Scarcity ended it. Markets ended it. Leasing ended it. Bankruptcy courts ended it. Security dependency ended it. Critical infrastructure ended it.

The question is no longer whether IPv4 has value. The question is whether private registries can keep exercising public-infrastructure power while denying property-like reliance, limiting liability, rejecting portability, and claiming community authority without lawful representation.

They cannot.

The way forward is not to make RIRs better sovereigns. The way forward is to end the sovereignty claim.

The registry layer should be thin. It should record uniqueness, control, transfers, security assertions, and conflict metadata. It should be auditable. It should be replaceable. It should have failover. It should not be able to destroy live network continuity because a private policy theory dislikes an operator’s business model.

Portability must become a hard right. If a registry becomes abusive, captured, insolvent, conflicted, or technically irrelevant, networks must be able to move their number resources and ASNs away from it. Without portability, every RIR is a lock-in point. With lock-in, every dispute becomes a hostage situation.

This is why the stability claim fails.

AFRINIC is not stability and Lu is not instability. That is propaganda geometry.

The real distinction is different. AFRINIC defends registry institutional stability. I defend routed network and customer-continuity stability. AFRINIC worries about the stability of its authority. I worry about the stability of the Internet that people actually use.

The Internet does not fail only when packets stop moving. It also fails when the institutions describing who may use identifiers lose the trust of the operators who move the packets.

That is the lesson.

Not mandate laundering.
Not running-code betrayal.
Running-code primacy.