Q5· lens: The stack

Where does durable competitive advantage settle, and does any layer earn outsized rents?

◇ Neutral question

Under-discussed relative to importance; the whole valuation case depends on someone capturing durable margin.

◷ Share view, as of June 2026: confidence, rationales and priors are hidden. The questions, stances, claims, tests and evidence stay visible.

Stances · candidate answers

Q5-S5A

Labs converge to commodity; advantage accrues to distribution and chips.

Held by analysts noting frontier models cluster tightly and API switching costs are low

Rents go to distribution (hyperscalers, OS / device) and scarce inputs (NVIDIA, TSMC, power).

would move me offA lab sustains a large, durable capability lead customers pay a premium for and won't switch from.

Strongest opposing: Q5-S5B

Evidence behind its claims: 187 support · 32 contradict · 95 neutral

Q5-S5B

Labs build durable moats via capability leads, ecosystems, capital scale.

Held by lab leadership, their investors

The capital to stay frontier is itself the moat; plus brand, safety reputation, ecosystem lock-in, data flywheels.

would move me offOpen-weight or cheaper models reach parity for the bulk of economic use cases, collapsing the premium.

Strongest opposing: Q5-S5A, Q5-S5C

Evidence behind its claims: 258 support · 35 contradict · 112 neutral

Q5-S5C

Open weights commoditize the middle; rents survive only at the extreme frontier and the physical layer.

Held by open-source proponents, some infra investors

Most value captured by good-enough open models running cheaply; labs fight over a shrinking premium tier while durable rents sit in silicon and power.

would move me offOpen models persistently lag far enough that enterprises won't deploy them for serious work.

Strongest opposing: Q5-S5B

Evidence behind its claims: 231 support · 26 contradict · 97 neutral

The map: how claims bear on stances

supports contradictshover a node to highlight its links

Claims

5.1

Frontier models cluster tightly (small gaps, fast catch-up).

Q5-S5AQ5-S5CQ5-S5B

5.2

Open-weight models reach "good enough" parity for most economic use cases.

Q5-S5CQ5-S5B

5.3

API switching costs are low (customers multi-home and swap easily).

Q5-S5AQ5-S5B

5.4

The capital scale to stay frontier is itself a sustainable barrier.

Q5-S5BQ5-S5AQ5-S5C

Stances

Q5-S5A

Labs converge to commodity; advantage accrues to distribution and chips.

held by analysts noting frontier models cluster tightly and API switching costs are low

Q5-S5B

Labs build durable moats via capability leads, ecosystems, capital scale.

held by lab leadership, their investors

Q5-S5C

Open weights commoditize the middle; rents survive only at the extreme frontier and the physical layer.

held by open-source proponents, some infra investors

Claims, each with its falsifiable test

5.1 · Frontier models cluster tightly (small gaps, fast catch-up).

would falsifyGap between #1 and #3 widens and persists across generations.

CapabilitySlow to resolve

5.2 · Open-weight models reach "good enough" parity for most economic use cases.

would falsifyOpen models stay materially behind on tasks enterprises pay for.

CapabilitySlow to resolve

5.3 · API switching costs are low (customers multi-home and swap easily).

would falsifyEnterprises show high lock-in, won't switch despite price / quality gaps.

MarketSlow to resolve

5.4 · The capital scale to stay frontier is itself a sustainable barrier.

would falsifyA new entrant reaches the frontier on materially less capital.

EconomicsSlow to resolve

Bridge-claims connecting out

⤧ B4Model commoditization determines where rent settles (away from labs).Capability → Market

Q5 / 4 claims / 3 stances / 1 bridge·confidence moves are made manually; AI proposes only