The Semi Impact Forum 2024 united investors, policymakers, and innovators from Taiwan and the UK in the semiconductor industry.
Held at the prestigious Royal Society on June 14 as part of the London Tech Week fringe event, the forum featured prominent Taiwanese figures such as Professor Konrad Young, former R&D Director at TSMC, and former UMC CEO Shih-Wei Sun. These experts shared their extensive experience and insights with the advisors and experts shaping the UK's Semiconductor Strategy, enabling an exchange between the leaders from both sides.
"Taiwan needs the world, just as the world needs Taiwan," says Representative Vincent Chin-Hsiang Yao of the Taipei Representative Office in the United Kingdom who opened the event with a speech.
"Semiconductors play a crucial role in international manufacturing and supply chains, driving technological breakthroughs in AI and other emerging technologies," Yao stated.
Yao highlighted Taiwan's leadership in the global semiconductor industry, anchored by three major science parks that serve as hubs for innovation, collaboration, and economic prosperity. He expressed Taiwan's readiness to collaborate with partners like the UK to drive future innovation and prosperity.
The wider landscape
On the morning panel, to give the UK audience less familiar with Taiwan's semiconductor landscape a better understanding, Konrad Young, the former TSMC R&D Director and now a professor at National Taiwan University split Taiwan's enabler platforms into three distinct parts to explain the ecosystem more succinctly.
"The first part consists of major companies such as TSMC and UMC, which are global leaders in the foundry sector," Young explained. "These companies are primarily business-oriented and ready to move to the next phase of production after Series B funding. Generally, such projects are difficult for the government or other entities to influence."
The second part, he illustrated involves more mature technologies and the early phases of pilot production. "Taiwan has several companies engaged in compound foundry work, capable of undertaking various tasks and collaborating internationally with the UK."
The third part, according to Young, is less common globally. "This phase involves very early-stage, deep-seed research and academic activities. Taiwan is currently developing infrastructure to support this phase, aiming to facilitate smooth transitions from academic research to early validation and then to production."
TSMC and UMC: the two sides of the story
In addition to mapping out the present-day semiconductor ecosystem, Young provided insights into the early challenges and growth of TSMC. He compared TSMC's humble beginnings to those of UMC and other pure manufacturing companies in the late 1980s.
"In 1987, TSMC emerged from ITRI (Industrial Technology Research Institute), while UMC had already been established for seven years," Young explained. "The intent was not to create a direct competitor to UMC. Instead, TSMC ventured into an area that no one else was interested in, largely due to a lack of confidence in the potential of a pure foundry model."
Young recalled how, despite having basic three-micron technology and suffering financial losses for nearly five years, TSMC gradually built trust with its customers. "Our early technology was rudimentary, but because no one else was doing it, we started to attract clients. TSMC's commitment to integrity and innovation—our slogan ICIC, standing for Integrity, Innovation, Commitment, and Customer Trust—was crucial."
Initially, TSMC operated as a pure manufacturing company, licensing technology from other companies such as Philips and manufacturing chips for clients such as Intel and AMD. "For the first 10 years, we didn't develop our technology. Instead, we focused on perfecting the manufacturing process and building trust with our customers," Young noted.
The turning point came when TSMC realized the need to develop its R&D capabilities. "When I joined TSMC, our technology lagged behind that of UMC. We decided to build an R&D team to catch up," Young said. "Despite the challenging environment, our persistent efforts and collaboration with experts from the US helped us overcome these obstacles."
Young also shared an antidote to illustrate the benefits of combining the best of both worlds from R&D and fabrication that took place under his tenure.
"When I returned to Taiwan, I aimed to create a 'smart R&D system' that combined the efficiency of manufacturing with the flexibility of research. This approach allowed us to rapidly iterate and innovate, making TSMC's R&D process unique and highly effective."
On the same panel as Young was UMC CEO Shih-Wei Sun, his former competitor who gave his side of the story during the early stages of these companies.
"UMC followed a different model as the first spin-off from ITRI," Sun explained. "Initially, UMC was an Integrated Device Manufacturer (IDM) with its products. We produced microprocessors and collaborated with companies like MediaTek, now a US$15-$20 billion company, and Novatek, the global leader in driver ICs."
Sun detailed UMC's transition from a product company to a pure foundry model. "We spun off our product companies, leaving the fabrication facilities to focus on manufacturing. This shift allowed us to specialize in wafer fabrication, progressing from six-inch to twelve-inch wafers."
However, Sun acknowledged challenges in UMC's technological advancements. "Our process technology faced setbacks at the 0.13-micron node. Unlike TSMC, which developed this technology in-house, UMC partnered with IBM in a joint venture that ultimately failed, nearly bringing UMC down. We had to recover from that."
Despite these challenges, UMC remained resilient. "Today, UMC is a US$10 billion company with a 30% gross margin and a 20% net profit, demonstrating our profitability and stability."
Sun also discussed the capital-intensive nature of the semiconductor industry. "The investment required is enormous, especially during the first depreciation cycle of five to ten years. It's a long and tough cycle, but once you get through it, the returns can be substantial."
He concluded by addressing the geopolitical landscape's impact on the semiconductor industry. "With regions worldwide aiming to develop their semiconductor manufacturing, it's crucial to be cautious and prepared for the long-term investment required. However, if managed well, the rewards can be significant."
The future
On the same note when asked if the UK's current direction of more selective investment into semiconductors is the right one especially when compared to the large-scale financial injection of the US, Young suggests the UK shouldn't follow what the other main players are doing, especially when they are already strong. Instead, they should focus on unique opportunities."
Young argues investing in the early phase of technology development can leverage future growth. "In the 1980s, the US invested heavily in semiconductors, but a smaller, well-placed investment can also yield significant returns," he added.
Regarding the future of the semiconductor ecosystem, Young believes education and AI will be essential not only to upscale production but also to evolve the entire system.
Young praised Taiwan's semiconductor education and suggested it could serve as a model for other countries. He believes Taiwan's semiconductor training is among the world's best. With decades of experience, this expertise could be translated to different cultural contexts to help other nations.
However, despite his praise for Taiwanese semiconductor training, he believes the system needs to cultivate more future leaders rather than basic workers. Hence why after retirement he started teaching semiconductor courses at the National Taiwan University to spearhead these initiatives
He rations that Taiwan's low birth rate and aging population necessitate a shift towards more innovative production systems, "a no-man fab," he suggests.
Young says his vision includes developing automated manufacturing systems. "The traditional working model won't work in the future. By creating systems with minimal human intervention, we can overcome cultural and logistical challenges."
Regarding talent, Young says he also observed a significant communication gap between Generation Z and baby boomers, emphasizing the need to bridge this divide. He praised Nvidia CEO Jensen Huang as a perfect example of someone who can connect with both generations, citing an instance where Huang took 90-year-old Morris Chang to a night market for the first time. Young added, "We need to recruit individuals with diverse educational backgrounds to help this industry grow, beyond just traditional technical expertise."
Furthermore, he underscored the role of AI in future fabrication processes. He believes AI should be a helper, enhancing human creativity and productivity.
"The production of semiconductors has matured so much from the physics to the working SOP, to the point that flow is now standard. So why do we need so many people working so hard just doing the tedious work when even the machine can do it?" Young questioned.
"You need to think, you need to design, you need to create," he says encouraging the next generation of talent.
When asked how he foresees the UK and Taiwan could help each other, Young sees the two tech powerhouses as a natural fit that complements each other's strengths and weaknesses. Taiwan can contribute to manufacturing and initial investment, while the UK excels in innovation.
"Your innovation will help us, bring us to the world. We work together to achieve high-value, high-profit, that's my recommendation," concluded Young to the British audience.
The event is supported by organizations such as Taiwan's National Applied Research Laboratories (NAR Labs), the National Science and Technology Council (NSTC), and the British Office Taipei, the event also saw participation from UK groups including techUK and Tech London Advocates & Global Tech Advocates.
Credit: DIGITIMES Asia
