Chen Liwu has been giving more interviews recently, with the latest episode being an interview with the NoPriors podcast. This time, the focus is on the reshaping of the semiconductor supply chain. His layout in advanced packaging and new materials fields, as well as his positioning in the foundry business, are worth paying attention to. He also revealed the progress of the Terafab project he is working with Musk (Intel provides technology and processes) and holds weekly meetings with him. He also discussed his semiconductor industry investment framework and promising new directions. Sort out the key points: 1. Chen Liwu clearly stated that the goal is to achieve a tenfold return for shareholders within 5-10 years. He has generated approximately 6 times the return for shareholders in the past 14 months, but emphasized that 'this is just the beginning'. He expects that the outside world will not truly see Intel's potential until 2030-2032. 2. The growth of AI demand faces several bottlenecks: firstly, power restrictions, as some countries simply do not have enough electricity; The second is the impact of helium gas, which many people are not aware of has a significant impact on the semiconductor industry; Thirdly, there is a shortage of memory, which is currently the most pressing issue 3. Technical roadmap reconstruction: Faced with the challenge of traditional process node miniaturization approaching physical limits, Chen Liwu is systematically shifting towards advanced packaging and new materials fields: 1) Promote EMIB advanced packaging technology and advance advanced packaging manufacturing projects in India and New Mexico, USA. 2) Invest in a glass substrate company (3DGS) and have confidence in the heat dissipation and insulation performance of glass. 3) Investing in a company that produces artificially synthesized diamond wafers, diamonds have enormous potential as insulation materials. 4) Layout in the field of new materials such as GaN (gallium nitride), SiC (silicon carbide), InP (indium phosphide), etc. 4. New opportunities brought by AI: The explosion of intelligent AI is driving a strong rebound in CPU demand. The ratio of CPU to GPU in data centers has decreased from 1:8 in the past to 1:4 or even lower. CPU performs better in reinforcement learning and agent coordination scheduling. Chen Liwu believes that AI will profoundly reshape the semiconductor industry, far beyond the influence of the Internet, and extend to edge computing, physical AI and other fields. 5. Positioning of foundry business: Adhere to the development of wafer foundry business, with the core logic being the strategic value of advanced manufacturing in the United States (supply chain security, not excessive concentration). The essence of OEM is a 'trusted business', focusing on yield, defect density, and cycle time. We are partners rather than mere competitors with TSMC. It is expected that the true potential of the industrial and commercial sector will only be revealed in the 2030-2032 period. 6. Co build Terafab project with Elon Musk 1) Joint judgment with Musk: The development of semiconductor infrastructure has not kept up with the growth of AI - there are gaps in production capacity, efficiency, and power consumption efficiency; 2) I really enjoy the process of working with Musk. He is very unconventional and always asks' why do we need to do it in a traditional way 'at every stage, which is very refreshing. I like to hear different opinions and then work together to find the optimal path, both sides can learn a lot. He has a clear vision - his robots and cars require a large number of chips. 3) Intel holds weekly meetings to provide technical and process support, and holds weekly meetings with Musk. 7. Chen Liwu's investment framework and optimistic direction for the semiconductor industry: 1) During the VC stage, Chen Liwu had a record of 159 IPOs and 126 M&A exits, with over 200 semiconductor investments and 38% in the United States. 2) In terms of investment methods, always start from a core question: where is the bottleneck and what problem are you solving. For example, I invested in Cradle Semiconductor because interconnection became a bottleneck; I invested in Celestial AI because optical interconnection is becoming increasingly important within clusters - it's not a coincidence that Huang Renxun has invested in almost all companies related to photonics. 3) The investment framework always remains: Is the problem real? Are customers really struggling with this? And most importantly, who is the first target customer? Tend to choose ultra large scale customers - they have the ability and willingness. If they like your product, they are willing to pay millions or even provide certain guarantees in the next few years, because after winning a large customer, they can scale up 4) Optimistic investment areas -There is a huge opportunity in the EDA field, can AI and machine learning help reduce complexity and improve design quality; -In terms of new materials, gallium nitride, silicon carbide, and indium phosphide are all investment directions; -Power management - the process of converting from 40V to 1V incurs significant losses - is also a promising bottleneck track; -Physical AI is the next major frontier, and we need to take a serious look at the entire stack. We are very optimistic about open-source cutting-edge technologies for physical AI, which is a gold mine. 8. Transformation framework and vision: Adopting a "crawling walking running" model, currently still in the "crawling" stage (consolidating foundation, building team). The long-term strategy is to integrate XPU, advanced packaging, and foundry capabilities to provide customized chip solutions for different workloads. It can be seen that this interview focuses on strategic transformation rather than short-term financial reports or stock prices.