Act III · Follow the bottleneck

The industry can survive while control moves to another layer.

Do not ask only whether an asset remains useful. Ask which interface, input or permission still constrains participation—and who owns it.

15 · Interface migration

A network and the interface that captures its value are different assets.

Profiles
Registries
Reputation
Marketplaces
→
Agent interface

Participants can remain in incumbent systems while discovery, comparison and intent move above them. Real counterparties and legal access resist this migration more strongly than information aggregation.

16 · The frontier race

Transforming a field does not transfer its historical expenditure to the transformer.

Frontier systems open new work.

new monetizable novelty ↑

Transformed work specializes and becomes cheaper.

mature capability cost ↓

Durable capture depends on which side of that race moves faster.

17–19 · Capital and substitution

Historical investment is not the cost of sufficient entry.

Incumbent

Accumulated plants, software, teams, workflows and distribution.

historical capital

Entrant

Current components required to deliver an adequate outcome by any route.

sufficient-entry capital

A breakthrough can be revolutionary and still face a price set by alternatives. High ROIC can remain available while becoming accessible to more producers.

20 · Three tests

A bottleneck must survive three different attacks.

Reproduce it
·
Substitute around it
·
Require less of it

Compute, energy, data, distribution, trust, networks, regulation and physical resources all face the same test. The one that resists becomes the candidate home of rent.

21 · The competing industrial shape

Concentrated infrastructure may sit beneath distributed productive agency.

Frontier dependence, vertical integration, regulation, trust or irreducible physical capital can preserve concentration. But concentration in one layer no longer proves that every downstream organization must remain large.