by Andhie Wong
In the fast-changing world of artificial intelligence, the United States and China are locked in a rivalry that goes well beyond technology — it touches economic power, ethics, and global influence. As Malaysia pushes forward with its own digital transformation, understanding this AI race offers real insight for our own technology goals.
Performance versus cost efficiency
US models like OpenAI's GPT-4 and Anthropic's Claude 3.5 Sonnet currently lead on performance benchmarks, scoring 84.5% and 83.7% respectively on tests like MMLU and HumanEval — a reflection of the strong, security-focused R&D systems behind them.
Chinese models like DeepSeek-V3 are closing the gap with a very different approach: lower training costs (around $5.5 million) and a shorter roughly six-month training cycle, while still landing solid MMLU and HumanEval scores of 82.3% and 83%. That efficiency shows how China is using resources to make advanced AI more affordable.
Different strengths, different markets
US and Chinese models reflect different priorities in their strengths and applications. US models are broadly versatile, powering conversational AI, API integrations and use cases across chatbots, content creation and research analytics.
Chinese models, by contrast, lean into multilingual capability, customer service, e-commerce and natural language processing, and are increasingly used in markets that value cost efficiency and ethically grounded AI.
Market dynamics
US firms tend to run higher-cost, quality- and security-focused operations, backed by a strong global ecosystem — the OpenAI API, partnerships with companies like Meta. Chinese companies like Alibaba, meanwhile, have significantly lowered LLM pricing, making advanced AI more accessible to a wider audience — a strategy that's strengthened their position across Asia while extending their global reach.
Economic and ethical stakes
The economic impact of this race is significant. US AI models are projected to contribute roughly $13 trillion to the global economy by 2030, driven by investment in generative AI startups and infrastructure that will create many jobs. Chinese models, meanwhile, are democratising AI access for SMEs by cutting costs, supporting smart manufacturing and automation.
Both sides are also making ethical progress in their own way — US models focus on continuous updates, bias mitigation and compliance with regulations like GDPR and CCPA, while Chinese models aim for ethical practice within local law, improving incrementally over time.
The geopolitical layer
Geopolitics runs through all of this. The US has imposed export controls on advanced AI technology aimed at slowing Chinese progress, but Chinese firms keep finding innovative ways to adapt and advance regardless — and are focusing on faster AI adoption across Asia, Africa and the Middle East as a result.
Looking ahead, US models will likely keep expanding API capability while prioritising safety and ethics in their research; Chinese models will likely keep strengthening multilingual support to build influence across Asian markets.
What this means for Malaysia
As Malaysia navigates its own digital transformation, watching this US-China rivalry play out offers real lessons. The US leads on performance; China leads on cost efficiency — both matter enormously for our SMEs.
By understanding these dynamics, Malaysia can position itself strategically to use AI for economic growth while promoting ethical innovation that benefits everyone. In this high-stakes race for global technology leadership, success won't come down to technical skill alone — it'll depend on building AI solutions that are inclusive, ethical and economically beneficial for every nation involved.
Part of the AK AI Corner column. Originally published in Oriental Daily (东方日报) on Mar 15, 2025.