Trend 1: AI is Fundamentally Redefining the Core Concepts of User Interfaces and Operating Systems

AI is no longer just a function; it's fundamentally changing how we interact with technology. Qualcomm CEO Cristiano Amon noted that, given that computers now have the ability to understand human languages, AI will phase out traditional user interfaces and become the center of the new mode of human-computer interaction. In the AI era, the operating system will essentially be elevated into a “productivity operating system”. The emergence of AI Agents is set to transform both operating systems and applications. No longer just a computing platform, the OS could evolve into a marketplace for these new agents.

AI Agents will be able to accurately understand the user’s intentions, access and integrate both structured and unstructured data, and assist users in completing a variety of tasks. These agents will continuously learn and adapt, delivering increasingly intelligent and personalized services. In enterprise environments, AI Agents will deeply integrate with legacy corporate systems or platforms to boost operational efficiency and productivity. Many companies are already exploring the use of dedicated AI Agents for specific business needs, such as request-for-quotation (RFQ) processing, payroll management, and survey administration.

This new model of AI-driven interaction and automation will significantly enhance productivity for both individuals and organizations. A prime example is the Microsoft–Qualcomm partnership on Copilot+ PC, with Qualcomm developing chips with sufficient AI processing power and Microsoft integrating AI into the operating system platform to enable the new paradigm of human-computer interaction. This is an exemplary demonstration of the new user interface and operating systems. 

Trend 2: Edge AI as the Key to Breakthroughs in AI Experiences

AI technology has made significant progress, and its development focus is shifting from data center training to real-world use cases and large-scale deployment. The AI computing capabilities of edge devices provide a strong foundation for the rollout of AI applications. A well-developed edge AI environment enables AI inference capabilities to be widely adopted by every device, i.e., this is the heart of large-scale edge AI use cases. 

Advancements in AI models are further accelerating the widespread application of edge AI. For example, with its relatively compact number of parameters, Deepseek achieves performance levels close to that of large language models such as GPT-4.0. This breakthrough signals that edge AI experiences are reaching new heights and offering users smarter and more efficient interaction. 

At COMPUTEX 2025, MediaTek showcased its prowess in pushing AI processing power to the edge. The flagship Dimensity 9400/9400 Plus chips can support over 240 AI models on smartphones, with deep inference capabilities built into the device to deliver an intuitive and natural AI Agent experience. MediaTek’s Genio IoT platform (e.g., Genio 720/520) supports the latest generative AI models and multimedia features, enabling devices to sense and respond to environmental signals in real time.

Trend 3: A Complete AI Ecosystem: From Data Centers to End Devices to Real-World Deployment

AI is fundamentally reshaping the technology industry at an unprecedented speed and depth. Its influence has extended beyond the cloud and is gradually spreading to edge devices. In addition to the focus on computing performance, the industry is placing great importance on real-time responsiveness, reduced latency, and the critical importance of user privacy protection. 

In terms of AI solutions for data centers, NVIDIA and server vendors jointly presented their latest AI compute “scale-out” offerings at COMPUTEX. By integrating optical fiber communication technology with co-packaged optics (CPO) chips, they have achieved an aggregated data transmission capacity of 400 Tb/s. 

At the enterprise level, major chipmakers have introduced their own AI compute solutions: NVIDIA launched GPU expansion cards for the RTX PRO Blackwell series; AMD introduced the Radeon AI PRO R9700 GPU; Intel unveiled the Arc Pro B50/B60 GPUs. These significantly enhance the AI capabilities of enterprise servers and greatly boost on-premises AI computing performance for corporates.

As far as personal devices are concerned, AI laptops and AI desktops were the key highlights. Qualcomm continues to focus on the WoA (Windows on Arm) AI PC market and announced it will launch its next-generation ARM PC processor in September. Intel and AMD demonstrated the capabilities of their AI PC chips to run AI applications on personal devices. MediaTek displayed its AI chips and solutions for the Google Chromebook system. The NVIDIA DGX Spark, which the company calls its highest-performance personal supercomputer, was presented by various PC brands each offering their own different hardware designs.

Implications for Taiwan’s Industries

The megatrends in the deployment of AI applications bring three major impacts to Taiwan’s semiconductor industry. First, the surging demand for AI chips — particularly driven by the AI-enabled transformation of user interfaces and operating systems as well as the widespread adoption of edge AI — has made advanced processes and leading-edge packaging technologies the key growth engines for the industry. This will continue to strengthen Taiwan’s leading position in semiconductor foundry and advanced packaging worldwide. 

Secondly, edge AI chips and solutions are emerging as a new market opportunity for the semiconductor industry. The demand for deploying AI inference capabilities to various end devices offers Taiwan’s IC design companies substantial growth potential, as these companies can leverage their existing advantages to develop high-performance, low-power-consumption chips dedicated to edge AI. 

Finally, the full-scale development of the AI ecosystem — from data centers to end applications — underscores the importance of heterogeneous integration and collaboration across the industry chain. This means Taiwan’s semiconductor companies must deepen their cooperation with global AI chip leaders and system integrators, and actively engage in the development of emerging technologies to ensure their pivotal position in the AI supply chain and enhance the value of their end-to-end solutions.