Beyond chatbots, two emerging AI races are underway—and they’re not being fought on the same turf

The AI landscape of 2026 is no longer a single pack. Two parallel dynamics are competing with different victory criteria—and conflating them distorts the analysis.

The Facts. Tech in Asia's analysis identifies two distinct dynamics in the evolution of AI: the race for foundational models (performance frontier, standardized benchmarks) and the race for deployed applications (market adoption, ROI, use cases). These two trajectories only partially overlap—and their leaders are not the same.

Our Take. The first race is being played out between five labs on codified benchmarks (MMLU, HumanEval, AA-Briefcase). The second is unfolding in real markets—South-East Asia, India, Korea, China—where the stakes are not frontier scores but commercial conversion and integration into existing workflows. Qwen and Kimi K3 are advancing less in U.S. rankings than in enterprise adoption across Asia. Models that win the first race do not necessarily win the second.

This framework has a direct implication for decision-makers: a model ranked #3 on a benchmark may be the right choice for a specific market if its team, infrastructure, and training data align with that context.

To Watch. The emergence of benchmarks tailored to the second race—measuring real-world adoption, not lab performance.