by Tan Aik Keong (AK)
For the past decade, the tech industry has measured breakthroughs in years. The PC, the internet, the smartphone, the cloud — each revolution took years to gradually reshape global industry.
AI is breaking that pattern. AI competition has entered an unprecedented high-speed era: a model considered the world's most advanced a month ago can face a serious new challenger within weeks.
Lay the AI developments of July and early August 2026 out on a timeline, and it's a genuinely striking model race.
In early July, Fable 5 re-entered the global AI spotlight, representing a new generation of frontier models pushing on reasoning, multimodal capability and autonomous task execution. On July 9, OpenAI released ChatGPT 5.6, pushing generative AI capability forward again — with the focus shifting from simply answering questions to helping businesses complete longer, more complex workflows. On July 16, during the World Artificial Intelligence Conference (WAIC), China's Moonshot AI launched Kimi K3, drawing global attention and signalling that Chinese AI companies are now genuinely competing at the frontier — closing the gap on long-context understanding, coding ability and AI agent task execution. By early August, Alibaba released Qwen 3.8, intensifying the competition further and reinforcing China's push toward an open model ecosystem that lets more developers and businesses build their own AI applications on top of advanced models.
Multiple major shifts, inside a single month. This is no longer competition measured in years — it's now measured in months, even weeks.
From chatbots to AI agents
When ChatGPT first appeared in 2022, it was most people's first real encounter with generative AI at scale. Back then, businesses were asking: can AI help me write? Can it answer customer questions? Can it make my staff more productive?
A few years later, the questions have completely changed. Now businesses are asking: can AI become a digital employee? Can it automate our business processes? Can it help us redesign how we work?
The underlying shift is that AI has evolved from a simple chat tool into an intelligent system with reasoning, tool use, coding, data analysis and autonomous task execution. The competition ahead isn't just about which model answers more accurately — it's about which one can complete more complex goals. Businesses used to ask AI to "write me a market analysis report." Now they want AI to gather market data, analyse competitors, organise customer feedback, produce a business briefing, and propose the next strategic move. That's the era of Agentic AI.
The US and China are now competing on two tracks
For the past few years, US AI companies have held a clear lead in foundation models. OpenAI, Anthropic and others have pushed the frontier through closed-source models: strongest model → cloud service → enterprise application. Businesses don't need to train their own models — they access the world's most advanced AI directly through an API or enterprise platform.
At the same time, Chinese AI companies are building a different path. The rapid progress of Kimi K3 and the Qwen series shows Chinese companies accelerating AI adoption through open models and developer ecosystems — a dynamic not unlike the smartphone industry, where Apple built the best experience on a closed ecosystem while Android let far more manufacturers and developers participate through an open one. The AI market may end up shaped similarly: a handful of companies holding the most advanced closed models, while a much larger number of businesses build their own AI solutions on open ones.
The real competition is AI infrastructure, not any single model
Many businesses fall into the same trap when they think about AI: constantly searching for "the world's best model." But the reality is that models are being updated faster and faster — whatever leads today may be overtaken again within months.
What businesses actually need to build isn't dependence on one model, but a set of AI capabilities that can adapt as things change. That means: model management — the ability to choose between models flexibly based on cost, performance and security needs; an AI Gateway to centrally manage different models with access control, cost management, data security and governance; an enterprise knowledge base, so AI understands your own data, processes and business experience; AI agents that don't just answer questions but actually execute work; and systems integration, connecting AI to your ERP, CRM, customer service and HR systems so it becomes part of how the business actually runs. Future competitiveness won't come from who has a ChatGPT account — it will come from who has built a complete AI productivity stack.
Malaysian businesses can't afford to wait for "the best model"
For Malaysian businesses, the biggest challenge AI's rapid pace creates isn't choosing which model to use — because that answer keeps changing. You might use OpenAI today and find another model fits certain use cases better tomorrow.
What actually matters is building model-agnostic AI capability. Just as businesses don't stop building mobile apps because iPhone, Samsung and other phone brands keep competing, businesses shouldn't wait for "the strongest AI" to appear before they start. The businesses that lead in future will be the ones that integrate AI into their workflows fastest, train their people on AI fastest, adopt AI agent ways of working fastest, and turn AI into business value fastest.
Closing: in the AI era, speed is the competitive advantage
Past tech revolutions took a decade or more to reshape the world. AI is moving at a completely different pace. In a single month, Fable 5, ChatGPT 5.6, Kimi K3 and Qwen 3.8 all launched in succession, reshaping the global AI competitive landscape each time.
This is only the beginning. What will ultimately decide the winners isn't who owns the biggest model — it's who can turn AI into productivity, business value and industry influence the fastest. In the AI era, speed itself is the biggest competitive advantage.
Part of the AK AI Corner column. Originally published in Oriental Daily (东方日报) on Aug 13, 2026.