ByteDance used AI that learns across products and an adaptive organizational structure to scale and diversify at once, reversing the usual trade-off between specialization and expansion; diversification amplified the value of its AI assets rather than diluting them.

Scaling high and wide: How firms leverage AI and organizational design to overcome the scale‐scope trade‐off

Feng Wan, Tianxi Yang, Xianwei Shi, Ke Rong, S. Ansari · Fetched April 25, 2026 · Strategic Management Journal

semantic_scholar descriptive low evidence 7/10 relevance DOI Source

ByteDance leveraged cross-domain AI learning and an adaptive organizational design to scale and diversify simultaneously, turning the traditional scale–scope trade-off into a strategic advantage.

The trade‐off between scale and scope has long posed a strategic dilemma, especially in digital settings, where specialization enables hyperscaling. Drawing on a longitudinal case study of ByteDance, we theorize how digital firms can overcome this constraint through the use of artificial intelligence (AI) combined with an adaptive organizational design. AI evolves and improves through self‐learning and cross‐fertilization across domains, becoming increasingly valuable as learning accumulates. This, however, is contingent on access to structurally related data that allow learning to transfer across domains. We show how AI reverses the conventional logic of the resource‐based view: rather than valuable resources enabling diversification, diversification amplifies the value of resources. AI thus transforms the scale‐scope nexus from being a trade‐off into a source of strategic advantage. The growing centrality of AI and digital platforms is reshaping how firms pursue and sustain growth. This study examines how ByteDance leveraged AI and adaptive organizational design not only to scale rapidly but also to diversify across industries and markets. Rather than incurring rising costs or coordination complexity, the firm's AI capabilities improved with each deployment through cross‐fertilization across domains, enabling more efficient growth across multiple domains. For managers, the findings highlight how dynamic combinations of AI and organizational structure can help overcome traditional trade‐offs between scale and scope, opening new pathways for scalable, cross‐market expansion in increasingly competitive environments.

Summary

Main Finding

AI, when combined with an adaptive organizational design, lets digital firms escape the traditional scale–scope trade-off. Through self‑learning and cross‑domain transfer of structurally related data, AI capabilities accumulate value with each deployment. Rather than diversification being enabled by valuable resources, diversification itself amplifies the value of AI resources, turning scale and scope into mutually reinforcing sources of strategic advantage. ByteDance’s longitudinal case illustrates this dynamic.

Key Points

Traditional dilemma: scale favors specialization (hyperscaling) while diversification increases coordination costs and dilutes scale advantages.
AI changes the calculus: models improve with continued use and by ingesting related data from multiple domains (cross‑fertilization), so widening scope can increase model value.
Critical condition: transfer requires structurally related data across domains; unrelated data yield less transferable learning.
Organizational design matters: adaptive, modular structures and coordination mechanisms enable rapid re‑use of AI across products and markets without prohibitive costs.
Reverse resource logic: instead of “valuable resources enable diversification,” diversification (when data and tasks are related) enhances resource value (AI models and data assets).
ByteDance example: the firm deployed a dynamic combination of AI and organizational adaptation to scale rapidly while entering new domains, improving AI performance with each new deployment.

Data & Methods

Empirical basis: a longitudinal case study of ByteDance (multi‑period examination of firm strategy and product evolution).
Analytical approach (as reported): process tracing of how AI capabilities were developed and reused across product lines, linked to organizational changes that supported cross‑domain learning.
Evidence types likely used in the study: time‑series of product launches and AI deployments, internal strategy patterns, descriptions of organizational adaptations, and performance trajectories showing cumulative learning effects. (The summary above reflects the study’s reported design; consult the paper for exact data sources and coding/validation procedures.)

Implications for AI Economics

Theory: AI reshapes firm boundary and resource theories — data and models are nonrival and accumulate value through related diversification, altering returns to scale and scope.
Market structure: platforms that can aggregate structurally related data across domains gain persistent competitive advantages; cross‑domain data complementarities become an axis of market power.
Strategy: firms should assess the structural relatedness of potential domains before diversifying; where relatedness exists, expanding scope can be the optimal path to amplify AI asset value.
Organization design: investments in flexible, modular structures, interoperable data pipelines, and governance that permits controlled cross‑use of data accelerate learning benefits.
Policy: competition and data governance policy should consider the amplification effects of cross‑domain data use—antitrust and data access rules may need to account for how related‑data aggregation generates durable advantages.
Measurement & research agenda: economists should refine models of nonrival learning spillovers, quantify relatedness between domains, and study welfare effects of AI‑driven diversification on entry, innovation, and consumer surplus.

Assessment

Paper Typedescriptive Evidence Strengthlow — Single-firm longitudinal case study that provides rich, plausibility-based inference but lacks counterfactuals, statistical tests, or causal identification; alternative explanations and selection/endogeneity cannot be ruled out. Methods Rigormedium — Likely rigorous qualitative methods (longitudinal tracing, triangulation of firm history, product deployments, and organizational changes) that support detailed theory-building, but the study does not employ quasi-experimental or quantitative causal methods and is vulnerable to single-case bias and limited transparency about data and coding. SampleA longitudinal, single-firm case study of ByteDance tracing its AI development, product deployments, and organizational adaptations over time using firm history, product and deployment timelines, public reports and filings, media accounts, and (presumably) interviews and internal documents or secondary sources. Themesorg_design innovation productivity GeneralizabilitySingle-firm evidence — findings may reflect ByteDance-specific capabilities, strategy, and culture rather than general patterns, Platform- and data-rich business model — results may not apply to firms without large, structurally related datasets, China-specific regulatory, market, and competitive context may limit transferability to other countries, Rapidly evolving AI technologies mean findings may be time-bound, Case focuses on successful firm — survivorship and selection bias limit applicability to less-successful firms

Claims (6)

Claim	Direction	Confidence	Outcome	Details
ByteDance leveraged AI and adaptive organizational design to scale rapidly and diversify across industries and markets without incurring rising costs or coordination complexity. Firm Productivity	positive	high	ability to scale and diversify across industries and markets (growth and diversification without rising costs/coordination complexity)	n=1 0.18
AI evolves and improves through self-learning and cross-fertilization across domains, becoming increasingly valuable as learning accumulates. Innovation Output	positive	high	AI capability improvement/value accumulation over time	n=1 0.18
The value of AI learning transfer across domains is contingent on access to structurally related data that allow learning to transfer across domains. Organizational Efficiency	positive	high	effectiveness of transfer learning across domains (dependence on structurally related data)	n=1 0.18
AI reverses the conventional logic of the resource-based view: rather than valuable resources enabling diversification, diversification amplifies the value of resources. Innovation Output	positive	high	amplification of resource value as a result of diversification	n=1 0.03
AI transforms the scale–scope nexus from being a trade-off into a source of strategic advantage. Firm Productivity	positive	high	ability to simultaneously achieve scale and scope (strategic advantage from combining both)	n=1 0.18
Dynamic combinations of AI and organizational structure can help managers overcome traditional trade-offs between scale and scope, opening pathways for scalable, cross-market expansion. Organizational Efficiency	positive	high	managerial ability to overcome scale–scope trade-offs and enable cross-market expansion	n=1 0.03