Suppliers can profitably abandon MCN reseller channels and switch to self-operated AI live-streaming when traffic spillovers and consumer preference for AI are strong, reshaping pricing and profits and — under some parameter combinations — boosting consumer welfare; MCNs may only gain when consumer AI preference is already high.

Citation: Hou W. & Chen N. (2026). Cooperation or AI-driven self-reliance? Supplier channel exit decisions in live‑streaming supply chain. Frontiers in Artificial Intelligence 9:1928289. doi:10.3389/frai.2026.1928289

Main Finding

A supplier’s decision to exit an MCN resale channel and operate solely via its own AI live‑streamer depends critically on two forces: the cross‑channel traffic spillover from MCN → supplier (θ) and consumers’ preference for AI streamers (k). Using a two‑stage Stackelberg model, the authors show that (1) channel exit can break the prior zero‑sum structure between supplier and MCN and produces region‑dependent equilibria; (2) under some parameter combinations a supplier exit can create a multi‑party win (supplier, MCN and consumers), but (3) strong consumer preference for AI frequently compresses consumer surplus and—counterintuitively—can improve MCN profits in some cases. The paper gives theoretical guidance on when suppliers should terminate MCN partnerships and how MCNs and regulators might respond.

Key Points

• Model setting: two consecutive selling stages. In stage 1 supplier and MCN operate in a dual‑channel mode (MCN resale + supplier AI channel). At the end of stage 1 supplier may revoke MCN authorization and switch to single‑channel AI monopoly in stage 2 (channel exit) or continue long‑term cooperation.
• Two main parameters drive outcomes:
  • θ: one‑way traffic spillover from MCN live rooms to the supplier’s AI live room (positive carryover).
  • k: multiplicative factor of consumer valuation for supplier’s AI channel (consumer preference for AI).
• Main comparative results:
  • Low-to-moderate k with sufficiently strong spillover (high θ) can produce Pareto improvements (higher supplier and MCN profits and higher consumer surplus) — a multi‑party win.
  • High k (strong AI preference) tends to reduce consumer surplus; in some parameter regions MCN profits increase even when the supplier focuses on its AI channel.
  • Channel exit alters pricing strategies and demand allocations; spillover mitigates the MCN’s exclusive bargaining power because supplier can internalize traffic gains.
• Managerial and regulatory takeaways:
  • Suppliers should consider exiting when own AI channel valuation (k) and accumulated spillover (θ) make self‑operation more profitable.
  • MCNs should account for spillover risk and may respond by changing effort levels, renegotiating exclusivity/commission terms, or adopting cooperative safeguards.
  • Regulators should monitor welfare impacts because AI adoption can compress consumer surplus even as some configurations increase aggregate welfare.

Data & Methods

• Type: theoretical (analytical) paper with numerical simulation for illustration and robustness checks.
• Game framework: two‑stage Stackelberg game between supplier (leader in certain choices) and MCN, solved by backward induction to obtain subgame perfect equilibria under two scenarios (long‑term cooperation vs short‑term cooperation/exit).
• Demand and preferences:
  • Consumer valuations v (MCN channel) and k·v (AI channel) are uniformly distributed on (0,1); k ≥ 0 indexes AI preference.
  • θ captures positive one‑way spillover from MCN stage‑1 traffic to supplier stage‑2 demand.
• Cost structure and assumptions:
  • MCN exerts effort e with quadratic cost 1/2 e^2.
  • Supplier production cost normalized to zero.
  • Supplier controls authorization of MCN resale channel (can revoke).
  • Interior equilibrium requires θ within specified bounds (ensures nonnegative demands).
• Outputs analyzed: equilibrium prices, demands, profits (supplier and MCN), consumer surplus and social welfare across parameter regions.
• Limitations documented by authors: stylized assumptions (uniform valuations, zero production cost, single MCN, two stages), no explicit modeling of commission contracts or multiple competing MCNs, and no direct empirical estimation.

Implications for AI Economics

• Platform and channel power: AI‑driven self‑operation changes bargaining leverage across the ecosystem because firms can internalize spillover traffic and data — altering platform/MCN business models and bargaining positions.
• Data ownership and strategic exit: control over user traffic and data (through AI channels) becomes a valuable strategic asset; suppliers may forgo intermediaries to capture data rents and downstream margins, reshaping industry structure.
• Welfare tradeoffs of AI adoption: AI streamers can improve operational efficiency and aggregate welfare in some regimes (strong spillover + moderate AI preference) but may compress consumer surplus when consumer AI preference is high — implying nonlinear welfare effects from technological adoption.
• Policy considerations: regulators should track how AI‑enabled channel consolidation affects consumer surplus, competition, and data concentration; interventions (e.g., transparency, data portability or limits on exclusive revocation) may be warranted depending on market outcomes.
• Research directions: empirical testing of model predictions (e.g., measured θ and k in real markets), extensions to multiple competing MCNs/platforms, endogenous commission contracts or exclusivity clauses, richer consumer heterogeneity, and multi‑period dynamics beyond two stages.

Limitations to keep in mind when applying the results: the model is analytic and assumes simplified cost and preference structures; real markets feature heterogeneous consumers, multiple intermediaries, and dynamic contractual arrangements that could modify thresholds for profitable exit.