This article is based on the talk given by Dr. Lisa Su, CEO of AMD, shared on YouTube, which delves into AMD's transformation strategy, the future trends of the AI industry, and her unique insights on leadership, innovation, and the semiconductor ecosystem in Taiwan. (Background: NVIDIA's RTX 5090 graphics cards are in short supply globally! There’s a rush to buy them at original prices, and even at double the price in the second-hand market, they are hard to find.) (Context: Jensen Huang announced the RTX 5090 graphics card with 'twice the performance of 4090', equipped with AI technology DLSS 4; NVIDIA's market value surged, overtaking Apple to become number one again.) Dr. Lisa Su, CEO of AMD and board member of the American Semiconductor Association, was invited to speak at National Taiwan University this month. She shared her journey leading AMD from adversity to success, her insights into the vast AI market, and her high regard for the semiconductor ecosystem in Taiwan. This semiconductor giant, born in Tainan and a PhD graduate in Electrical Engineering and Computer Science from MIT, brought a deep conversation about technology, strategy, and life philosophy to the students and industry professionals present. Here are the key points compiled by us: At the beginning, Lisa Su humbly expressed that interacting with so many talented students made her 'feel old', but she hoped to share some stories to inspire the students' careers. She reflected on her academic background, from her bachelor's, master's, to PhD at MIT, specializing in semiconductor devices. Over thirty years of experience in the semiconductor industry has given her a profound understanding of the rapid development of technology. Since joining AMD in 2012 and taking over as CEO in 2014, she has established a clear position for the company: focusing on 'High-Performance Computing' (HPC). 'High-Performance Computing has essentially become the core of our lives today.' Lisa Su emphasized that AMD's technology reaches billions of people every day, whether in cloud data centers, network edges, personal computers, automobiles, or industrial sectors, one can see the presence of AMD technology. She pointed out that as a technology company, it is crucial to have a long-term strategy and technology roadmap. AMD's success is based on three pillars: Focusing on high-performance computing: Making HPC the core competency and continuously investing in R&D. Establishing excellent partnerships: TSMC is one of AMD's core partners in co-developing leading products. Embracing technological change and betting on Chiplet technology: Given the slowdown of Moore's Law, AMD boldly bets on Chiplet technology, seeing it as a key to future development. These strategies have shown significant effectiveness, leading Lisa Su to proudly state that currently, five of the top ten supercomputers in the world utilize AMD technology. 'I firmly believe that high-performance computing will be the most important core area in the future.' she said. The AI wave: A $500 billion market and overall design thinking When discussing the hottest AI topics, Lisa Su admitted that the pace of innovation in the AI industry is astonishing, 'almost new technologies and innovations emerge every week.' She cited the DeepSeek model that drew attention this January, explaining how it successfully lowered the barriers and costs of AI usage through techniques such as the Mixture of Experts, thereby accelerating the popularization and application of AI. 'The main contribution of DeepSeek is that it has truly accelerated the application of AI.' Lisa Su emphasized that there is no 'one-size-fits-all' solution in the AI computing ecosystem. On one hand, it requires hundreds of billions of dollars to build the most powerful frontier large models; on the other hand, there is also a significant demand for numerous medium and small models, as well as various algorithms. She anticipates that investment in the AI field will continue to grow significantly in the future, covering both large and more cost-effective investments. 'Regarding AI, the most important point is: the market size is extraordinarily vast.' Lisa Su predicts that within the next three to four years, the AI market size will reach $500 billion. Such a vast market requires various computing capabilities, including CPU, GPU, ASIC, FPGA, etc. 'No single company can provide all solutions.' She believes that AMD's AI strategy consists of two main points: Providing complete solutions: Covering hardware, software to system levels. Establishing an open ecosystem: Encouraging collaboration and co-innovation, allowing everyone to learn from each other. She again mentioned the open model characteristics of DeepSeek, which promotes secondary innovation based on its foundation. Lisa Su further pointed out that AI has permeated every layer of AMD's business, from the largest data centers to personal computers, there will be a large number of AI applications. 'This means a tremendous market opportunity for all of us.' She also reminded that especially in the AI field, design thinking must change. 'You must conduct holistic design.' She stressed that hardware, software, and systems must be designed collaboratively; they can no longer be considered separately as in the past; one must start from the overall system applications and algorithms to rethink how to design chips. The key role and resilience of the Taiwanese ecosystem Lisa Su highly values the semiconductor ecosystem in Taiwan. 'Taiwan's ecosystem is very special.' She pointed out that from TSMC's excellent manufacturing capabilities to the complete industry chain covering ODM, OEM, manufacturing, and design, it is impressive. She expects that future investments in software and systems will increase. AMD's team in Taiwan is an essential part of its global R&D, and Lisa Su is very proud of this. At the same time, she emphasized the interdependence of today's global ecosystems. 'From the perspective of the United States, we indeed rely heavily on Taiwan's manufacturing ecosystem.' Therefore, establishing 'resilience' for this ecosystem is crucial. This also echoes the recent global supply chain restructuring and diversification trends, highlighting Taiwan's indispensable position in the global technology industry. This CEO, who was born in Tainan and later immigrated to the United States, also shared her personal growth experiences. She thanked her parents for emphasizing education; even when she was eager to enter the workforce, her parents still 'strongly suggested' that she pursue a PhD. 'It turns out my parents were right.' She believes that the experiences in college and graduate school provided a valuable opportunity to learn diverse knowledge and build confidence to solve future problems. From a semiconductor engineer to a business operator, and then to a semiconductor leader, this journey of 'dream job' has not been easy, especially in the early days, family support was crucial. Leadership, innovation, and advice for students Lisa Su believes that leadership styles vary, but university education is very helpful in nurturing leadership. She shared one of the best pieces of advice she received: 'Go solve the hardest problems.' She encourages students to choose projects, problems, or companies that can have a significant impact because they can learn the most and contribute the greatest. 'One thing is for sure, I learned the most from the biggest mistakes I made.' She believes that mistakes are not bad; they are opportunities for learning and improvement. Regarding AMD's transformation journey, she recalled the difficulties she faced when she first took over in 2014. At that time, the company's product roadmap lacked competitiveness. 'In a tech company, products are everything.' Therefore, her top priority was to reorganize the product roadmap and establish three main priorities: creating excellent products, building deep customer relationships (emphasizing the importance of collaboration), and maintaining agility and speed of action.

#AI