csQCA · Working Paper · 2026

Business Model Archetypes in AI Foundation Models

A crisp-set Qualitative Comparative Analysis of N=49 AI foundation model vendors, examining which configurations of market position, integration strategy, and openness are sufficient for commercial viability.

0
Vendors
0
Configurations
0
Solution Pathways

AI Foundation Models

Foundation Models (FMs) are large neural networks trained on broad data at scale and adaptable to a wide range of downstream tasks (Bommasani et al., 2021). Two defining properties distinguish them from narrow AI: emergence—qualitatively new capabilities arising at scale—and homogenization—convergence of diverse AI applications toward a shared representational substrate.

4 Structural Economic Features (Burkhardt et al., 2024)
1
Capital-Intensive Pre-Training
Frontier model training exceeds $100M per run, creating structural barriers to entry.
2
Economies of Scale in Inference
Marginal cost per query falls sharply at scale — cost advantage compounds with volume.
3
Data Network Effects (RLHF)
User interaction signals feed back into model quality, compounding advantage for large-base vendors.
4
Non-Rival Knowledge Accumulation
Trained weights serve unlimited concurrent users at near-zero marginal cost after fixed training.
4 Vendor Categories in This Market
Proprietary ClosedRetain weights, access via API subscriptions, accumulate exclusive RLHF signal.
Open-Weight CommunityRelease weights under permissive licenses; monetize through ecosystem and complementary products.
Platform-EmbeddedIntegrate AI into existing ecosystems (Search, Social, OS); monetize via advertising or subscriptions.
Specialized DomainApply FM capabilities to high-value verticals (legal, medical, financial); command premium pricing.

The coexistence and commercial viability of all four categories simultaneously constitutes the central empirical puzzle motivating this study.

Key Finding

~INTEGRATED + MONOPOLY·WIDE → Y=1

Overall consistency = 0.917 · Coverage = 1.000 · N=44 viable vendors

Term 1 — Collective Path
~INTEGRATED
cons=0.919 · cov=0.773 · n=34 vendors
Term 2 — Proprietary Path
MONOPOLY·WIDE
cons=0.929 · cov=0.295 · n=13 vendors

Theoretical Framework — PEMAK

Brousseau & Pénard (2007) identify three platform functions that apply to all digital platforms — and to AI foundation model vendors.

Matching (M)
User intent → model response. Conditions: MONOPOLY, INTEGRATED
Assembling (A)
Bundling AI across products. Conditions: WIDE, PAID
Knowledge Management (K)
RLHF pipeline governance. Conditions: HIERARCHICAL, OPEN
6 conditions → 64 possible combinations → 13 actually observed

Complementor Engagement (§2.4)

FM vendor viability depends on third-party complementors building on the model. Two distinct mechanisms achieve this threshold.

Active Orchestration
Proprietary vendors (Anthropic, Cohere): SDK, developer programs, enterprise API — despite closed weights.
Proprietary Path
Ecosystem Generativity
Open-weight vendors (Meta, DeepSeek): permissive license → community complementors at near-zero cost (Zittrain, 2006).
Collective Path
Chanson-Rocchi structural analogue: CE = API adoption threshold preceding FM vendor viability. Construct content differs from PL (neobank market recognition); threshold logic is preserved.

Key Findings

Equifinality

P1 + P2

Two logically independent pathways — Collective Path (~INTEGRATED) and Proprietary Path (MONOPOLY·WIDE) — produce the same outcome of commercial viability through fundamentally distinct causal mechanisms.

Death Zone (n=0)

Structural Absence

No vendor achieves INTEGRATED=1 without MONOPOLY=1 or OPEN=1. Full vertical stack ownership without adjacent monopoly cross-subsidy is structurally prohibitive at frontier training costs.

OPEN as Insurance

Necessity (Y=0)

13/13 open-weight vendors (OPEN=1) achieved Y=1. OPEN=1 has Cov(Y=1)=1.000: open-weight release is a sufficient insurance against organizational exit in this sample.

No Single Necessary Condition

QCA Finding

No single condition meets the necessity threshold (cons≥0.90) for Y=1, confirming the equifinality hypothesis. Multiple distinct structural configurations achieve viability.

4 Theoretical Contributions

01

CE Threshold

Grounds FM viability in active 3rd-party adoption. The moment a technology project becomes a business platform.

02

Ecosystem Generativity Insurance

Open-weight release as structural insurance. 20/20 vendors viable — zero exceptions. Zittrain (2006) operationalized.

03

Functional Equivalence

FM vendors ARE platforms if they match, assemble, and manage knowledge. PEMAK extends to AI without modification.

04

Equifinality as Stable Equilibrium

Multiple paths to viability are NOT transitional. Capital constraints + non-fungible network effects sustain the ecology.

Methodology

Citation

Xu, J. (2026). Model of Models: A Configurational Analytical Framework of Business Models for AI Foundation Models. Working paper v6.9.