Lisa Su at NTU: AI, DeepSeek, and Taiwan’s Edge in Uncertain Times
Insights on AI, Semiconductors, and Leadership
Lisa Su, CEO and Chair of Advanced Micro Devices (AMD), returned to her roots in Taiwan for an inspiring and wide-ranging discussion on technology, leadership, and the future of AI at a university forum hosted by National Taiwan University (NTU) President Chen Wen-chang. Addressing a packed auditorium of students and faculty, Su offered not only a glimpse into the cutting edge of semiconductor innovation, but also her personal journey—from a curious MIT student to the helm of one of the world’s most influential semiconductor companies.
A Lifelong Passion for Semiconductors
Dr. Su began her talk by engaging with students directly, asking how many were studying electrical engineering or the medical sciences, before recounting her own path through MIT, where she earned her bachelor’s, master’s, and Ph.D. in electrical engineering and computer science, specializing in semiconductor devices. "Technology moves so fast," she noted, "and after more than 30 years in the semiconductor industry, I still feel we are only scratching the surface."
At AMD, which Su joined in 2012 and has led as CEO since 2014, she spearheaded the company’s focus on high-performance computing—a decision that has positioned AMD at the core of cloud computing, edge devices, PCs, automotive, and industrial applications. “Our technology touches billions of people every day,” Su emphasized.
Strategic Vision: Performance, Partnerships, Packaging
Su outlined AMD’s three-pronged strategy for success in a rapidly evolving landscape:
High-Performance Computing: Doubling down on compute performance across CPUs, GPUs, and accelerators.
Deep Partnerships: Especially with foundry leader TSMC, a vital ally in AMD’s manufacturing ecosystem.
Chiplet Innovation: A strategic bet on chiplet architecture, key to maintaining Moore’s Law in an era of slowing transistor scaling.
This strategy, Su said, has paid off. “Today, five of the top 10 supercomputers in the world are powered by AMD.”
AI’s Next Frontier: Holistic Design and Democratization
Asked about the most exciting developments in chip design, Su stressed a holistic approach. “You can’t design AI in silos. Hardware, software, and systems must be co-optimized, especially when considering end applications and algorithms.”
She also addressed the rise of DeepSeek, a Chinese-developed AI system that stunned the industry earlier this year. “They’ve done something clever—by combining multiple models efficiently, they made AI more accessible and cheaper, which accelerates adoption.”
Despite fears that such breakthroughs could derail Western AI dominance, Su expressed confidence in the continued investment. “The AI market is expected to hit $500 billion within four years. There's room for frontier models, medium-sized applications, and lightweight edge AI—all requiring different compute architectures.”
Competing in a Crowded AI Landscape
On how AMD differentiates itself from rivals like Nvidia, Su was pragmatic. “No one company can do it all. We believe in open ecosystems, and AMD’s approach is to offer a complete solution—hardware, software, and systems. We are deeply embedded in everything from the largest data centers to AI PCs.”
She also underscored the importance of open collaboration, referencing how open models like DeepSeek enable innovation beyond their original creators.
Taiwan’s Edge and Advice for the Next Generation
When asked how Taiwan can stay competitive in the global AI race, Su praised the island’s world-class semiconductor ecosystem, led by TSMC and its network of ODMs and OEMs. Her suggestion? “Invest more in software and systems to complement the hardware strength. And build resilience—because the global tech ecosystem is interconnected.”
Su, who immigrated to the U.S. as a child, also touched on the impact of education and family values. Recalling her parents' insistence she finish her Ph.D.—even when she felt impatient to join the workforce—Su shared, “I didn’t love it at the time, but they were right. That education gave me the tools and confidence to solve hard problems.”
Resilience, Leadership, and Bold Ambition
Fielding questions from students, Su shared candid reflections on AMD’s turnaround from its most difficult years. “When I joined, we were not competitive. We rebuilt our roadmap from scratch. I always tell students—run toward hard problems. That’s where the most learning and impact come from.”
In response to a question about balancing ambition with humility, especially for Taiwanese and Asian students, Su offered this advice: “Be very ambitious. Say what you want to achieve. People will help you get there. Sometimes we are shy—but it’s okay to dream big.”
Closing Thoughts: Impact at Scale
Dr. Su closed by reminding students of the privilege and responsibility of working in technology. “We’re lucky. We’re working on things that will improve lives—healthcare, sustainability, productivity. The world needs innovation. And it needs people like you.”
Other takeaways:
🧭 AMD’s Vision & Strategy
AMD’s long-term strategy centers around:
High-performance computing.
Strong partnerships, notably with TSMC.
Innovation in chiplet technology, especially as Moore’s Law slows.
AMD technology now powers 5 of the top 10 supercomputers globally, reflecting its leadership in HPC.
Lisa emphasizes that strategy and focus are more important than company size — having "the right people" matters more than having the most people.
💡 AI & the Future
AI is transforming the computing landscape, and there's no one-size-fits-all solution — CPUs, GPUs, ASICs, FPGAs all play a role.
AMD’s AI strategy is focused on:
A complete solution (hardware, software, systems).
Embracing an open ecosystem that fosters collaboration.
She notes that innovations like DeepSeek are clever and impactful, especially in making AI cheaper and more accessible.
AI will influence everything from productivity to healthcare, where it can help in disease diagnosis, drug discovery, and improving patient outcomes.
🌏 Global & Geopolitical Viewpoints
Lisa highlighted the importance of Taiwan’s semiconductor ecosystem, especially TSMC, and sees growing opportunities in software and system design.
On global AI regulation, she advocates for balanced oversight that protects society without stifling innovation.
She acknowledged that countries embracing AI innovation will reap long-term benefits.
📚 Advice to Students
Run toward hard problems — tackling tough challenges builds the most growth.
Education is crucial. Lisa credits her parents for encouraging her to get a PhD, which built her confidence and skills.
You don’t need to be an engineer to make an impact in AI — interdisciplinary applications, even in literature or political science, are growing.
Embrace continuous learning and be agile in thinking; the world is evolving rapidly.
Set big goals and don’t be afraid to dream ambitiously — and share those dreams so others can help support them.
Failure is okay — it’s often the best teacher.
🔥 Leadership & Culture
Leading AMD through difficult times meant:
Building great products.
Developing deep customer relationships.
Staying agile and moving fast.
Motivation comes from seeing progress and impact, not just from incentives like stock prices.
Leadership styles differ — Lisa encourages ambitious, thoughtful, and collaborative leadership, especially in Taiwanese and Asian cultures where people may be more reserved.