Technology (T):
Digital and intelligent infrastructure
Moving beyond a mere exogenous efficiency parameter, T explicitly models digital and intelligent infrastructure to define the usable service capacity delivered by networking, compute, and storage under energy constraints. GDII 2026 decomposes T into networking (N), compute (C), and storage (S). Rather than a simple linear aggregation, these three types of services are combined using a Constant Elasticity of Substitution (CES) function with low substitutability:
Data (D): High-quality data
D goes beyond data volume accumulation to unlock value, capturing the effective data value jointly determined by data scale, quality, and velocity. GDII 2026 decomposes D into three dimensions: It uses a CES function to aggregate private data (DP, t) and public/government data (DG, t), defining the complementary structure of these two data sources. It employs data quality (Qₜ) as a multiplicative scale factor, reflecting the surge in value density as a result of long-term accumulation, standardized governance, and scenario-specific immersion. It uses velocity (vₜ), which is the ratio of accessed data volume to effective data volume, to capture the efficiency of converting data from a static asset into a dynamic factor of production:
Unlocking the value of data depends not on how much data is owned, but on "how much is usable, how well it is used, and how fast it moves".
Labor (L): Talent ecosystem
L moves labor beyond quantitative aggregation to human-machine collaboration, capturing the complementary aggregation of high-skilled talent and universal labor moderated by structural talent matching. GDII 2026 decomposes L into two types of labor: high-skilled talent (Hₜ) which encompasses AI-related professionals, and universal labor (Uₜ) which represents the general workforce. The two are aggregated using a CES function with low substitutability:
In the era of the intelligent economy, the talent structure needs to be coordinated with digital and intelligent infrastructure and high-quality data, and the talent ecosystem must focus on cultivating interdisciplinary talent and developing human-machine collaboration models.
