AMD CEO: China 20% of Revenue, CPU Boom Coming

AMD CEO Lisa Su Confirms China's Critical Role Amidst AI Shift

AMD CEO Lisa Su has explicitly stated that Mainland China remains a vital component of the company's global strategy, accounting for approximately 20% of total revenue. This admission comes at a time when geopolitical tensions often overshadow technological cooperation between Western firms and Chinese markets.

Su made these remarks during an exclusive interview with Nikkei Asia, highlighting the sustained importance of the region despite broader industry shifts. She emphasized that AMD maintains substantial operations across multiple sectors within China, including personal computing, gaming, and specific data center applications.

The statement serves as a strategic reassurance to investors and partners regarding AMD's resilience and diversified market presence. Unlike some competitors who may be retreating from certain regions, AMD is doubling down on its established footprint.

Key Takeaways from AMD's Strategic Outlook

• China Revenue Share: Mainland China contributes roughly 20% of AMD's total global revenue.
• Market Segments: Significant business activity exists in PC, gaming, and partial data center sectors in China.
• CPU Growth Prediction: The CPU market is projected to grow at over 35% annually for the next five years.
• AI Driver: This growth is fueled by the rise of AI inference and agent-based AI workloads.
• Historical Context: Previous CPU growth was stagnant at just 3-4% annually, with focus heavily skewed toward GPUs.
• Recent Engagement: Su attended the AMD AI DevDay 2026 in Shanghai, marking the event's first launch in China.

The Unexpected Renaissance of the Central Processing Unit

For the past several years, the technology industry has been obsessed with Graphics Processing Units (GPUs). These chips have become the undisputed heroes of the generative AI boom, powering everything from large language model training to complex image generation tasks.

Consequently, the humble Central Processing Unit (CPU) was largely overlooked by investors and tech enthusiasts alike. Growth rates hovered around a modest 3-4% annually, reflecting a mature market that seemed to have reached its peak potential.

However, Lisa Su argues that this narrative is fundamentally changing. As AI models move from the training phase to the inference phase, the computational demands shift significantly. Inference requires high-throughput, low-latency processing that CPUs are uniquely positioned to handle efficiently.

Furthermore, the emergence of AI agents—autonomous software systems that can perform complex tasks—places new burdens on central processors. These agents require robust logic handling and memory management, areas where traditional CPUs excel compared to specialized accelerators.

Su predicts that this shift will drive the CPU market to expand at a rate exceeding 35% per year over the next half-decade. This projection represents a tenfold increase in growth velocity compared to recent historical trends.

This forecast challenges the prevailing assumption that AI hardware demand is exclusively tied to GPU architectures. It suggests a more balanced ecosystem where CPUs and GPUs work in tandem to deliver comprehensive AI solutions.

Why Inference Changes the Hardware Equation

Training AI models is computationally intensive but occurs relatively infrequently. In contrast, inference happens every time a user interacts with an AI application, making it a continuous and volume-heavy process.

CPUs offer greater flexibility and cost-effectiveness for these varied workloads. They can handle diverse instruction sets and integrate seamlessly with existing enterprise infrastructure without requiring massive architectural overhauls.

As companies seek to deploy AI at scale, the economic efficiency of using CPUs for inference becomes increasingly attractive. This trend is particularly relevant for enterprises looking to balance performance with operational costs.

Deepening Ties with the Chinese Technology Sector

AMD's commitment to China is not merely rhetorical; it is backed by tangible actions and significant investment. The company recently hosted the AMD AI DevDay 2026 in Shanghai, a major milestone for its engagement in the region.

This event marked the first time the global developer conference landed in China, signaling a strong desire to foster local innovation. Lisa Su delivered the opening keynote, underscoring AMD's long-term dedication to the Greater China market.

The presence of AMD leadership at such events helps build trust with local developers and enterprise clients. It demonstrates that the company is willing to engage directly with the community, despite the complex regulatory environment.

In the personal computer sector, AMD continues to compete fiercely against Intel, offering competitive Ryzen processors that appeal to both consumers and businesses. The gaming segment also remains a stronghold, with Radeon graphics cards maintaining a loyal user base in China.

Additionally, AMD is expanding its footprint in the data center arena. While competition is intense, the company's EPYC processors are gaining traction due to their performance-per-watt advantages.

Strategic Implications for Global Supply Chains

Maintaining a strong presence in China allows AMD to mitigate risks associated with supply chain disruptions. By diversifying its manufacturing and sales channels, the company ensures greater stability.

This approach contrasts with strategies that prioritize complete decoupling. Instead, AMD opts for a nuanced path that balances compliance with commercial opportunities.

Such a strategy requires careful navigation of export controls and international trade laws. However, it positions AMD as a reliable partner for Chinese firms seeking advanced semiconductor technology.

Industry Context: The Broader AI Hardware Landscape

The prediction of a 35% annual growth in the CPU market aligns with broader trends in enterprise IT spending. Companies are no longer just experimenting with AI; they are integrating it into core business processes.

This integration requires a holistic approach to hardware. While GPUs handle the heavy lifting of matrix multiplications, CPUs manage the orchestration, data preprocessing, and post-processing tasks.

Competitors like Intel and NVIDIA are also recognizing this synergy. NVIDIA's recent acquisitions and partnerships highlight a similar understanding that AI infrastructure is not solely about accelerators.

However, AMD's specific focus on the inference workload gives it a unique angle. By optimizing its CPU architecture for these tasks, the company can capture a larger share of the emerging AI edge market.

The shift also impacts software development. Developers must now write code that efficiently utilizes both CPU and GPU resources, leading to more sophisticated optimization techniques.

What This Means for Businesses and Developers

For enterprise leaders, Su's comments suggest that investing in modern CPU infrastructure is a prudent strategy for AI readiness. Upgrading to latest-generation processors can yield significant performance gains for inference workloads.

Developers should pay close attention to tools that optimize CPU-based AI execution. Frameworks that support hybrid CPU-GPU workflows will become increasingly important in the coming years.

Small and medium-sized businesses, which may lack the budget for extensive GPU clusters, can leverage powerful CPUs to deploy smaller, specialized AI models effectively.

This democratization of AI compute power could accelerate adoption across various industries, from healthcare to finance.

Looking Ahead: The Next Five Years

The next five years will likely see a convergence of CPU and GPU technologies. We may witness more integrated solutions that blur the lines between these distinct processor types.

AMD's roadmap will need to reflect this reality, ensuring that its products remain competitive in a rapidly evolving landscape. Continued investment in research and development will be crucial.

Global markets will continue to play a pivotal role in this expansion. Regions like China, India, and Southeast Asia will drive much of the demand for affordable yet powerful AI hardware.

Stakeholders should monitor AMD's product launches and strategic partnerships closely. These indicators will provide further insight into how the company plans to capitalize on the predicted CPU boom.

Ultimately, the message from Lisa Su is clear: AI is not just a GPU story. It is a comprehensive computing challenge that requires the full spectrum of silicon capabilities.