Intel CEO Chen Lifang’s First Podcast Interview: Our Goal Is “10x in 5–10 Years,” Betting on Advanced Packaging, Glass Substrates, and Artificial Diamonds

Source: Wall Street Insights

Intel CEO Chen Liwu stated that his return target for Intel is "10 times within 5 to 10 years," and he is systematically reconstructing Intel's technology roadmap around advanced packaging, new semiconductor materials, and next-generation substrate technologies.

In a recent podcast episode, Chen Liwu elaborated on his path to transforming Intel: after solidifying the balance sheet and focusing on product lines, he is shifting his investment focus to advanced packaging technology EMIB, glass substrates, and new materials such as gallium nitride (GaN), silicon carbide (SiC), indium phosphide (InP), and synthetic diamonds to address the challenges of shrinking traditional process nodes approaching physical limits. He also revealed that the explosive growth of AI and inference scenarios is driving a strong rebound in CPU demand, with the CPU-to-GPU ratio in data center servers evolving from about 1:8 to 1:4 or even lower.

Chen Liwu said that in the past 14 months, Intel has generated about 6 times the return for shareholders, but "this is just the beginning." He expects that by 2030 to 2032, the outside world will truly recognize Intel’s potential—not only in the traditional PC client market but also extending into emerging markets such as edge computing, physical AI, and intelligent agents AI.

In his view, if Intel’s XPU, advanced packaging, and foundry capabilities can be effectively integrated, it will provide customized chip solutions for different workloads, which is his long-term strategic direction for the company.

Key to Breakthrough: New Materials, Focus on Advanced Packaging and Glass Substrates

Against the backdrop of shrinking traditional process nodes approaching physical limits, Chen Liwu points to materials science and advanced packaging as the breakthrough areas. He stated that Intel has already mass-produced 18A process, is pushing for 14A mass production, and can see technical paths toward 10 nm and even 7 nm, but "this path will become increasingly expensive and difficult."

To this end, Chen Liwu has launched multiple initiatives in packaging materials. He invested in glass substrate company 3DGS, appreciating glass’s unique properties as a heat dissipation and insulation material; in inter-chip interconnects, Intel is promoting the next-generation advanced packaging technology EMIB and has announced cooperation projects in India and New Mexico, USA to advance packaging manufacturing. Intel holds about 1,000 patents in module technology, and effectively integrating substrates and modules remains a core engineering challenge he emphasizes.

In the field of new semiconductor materials, Chen Liwu said he has invested in GaN, SiC, and InP, some of which have been acquired by large semiconductor companies like ADI. He also invested in a synthetic diamond wafer company, optimistic about diamond’s potential as an insulating material in chip packaging. "The engineer’s spirit is like this—constantly encountering bottlenecks, then finding ways to overcome or bypass them," he said.

Foundry Business: Trust First, Yield and Cycle Time Are Core Metrics

Intel’s wafer foundry business was once considered unsustainable by outsiders, but Chen Liwu chose to persevere. He explained that the core logic behind this decision is that advanced manufacturing in the U.S. has strategic value for supply chain security; no large semiconductor company can concentrate its supply chain in just one or two regions.

Operationally, Chen Liwu prioritizes yield, defect density, and cycle time for the foundry business. He emphasized that foundry is fundamentally a trust business—"before customers hand over wafers, they must trust you." If yield doesn’t meet standards, customers will lose revenue and leave, which is hard to recover from.

He also stated that Intel’s partnership with TSMC is not purely competitive; the industry as a whole needs more capacity to meet ongoing demand. He expects that by 2030 to 2032, the true potential of Intel’s foundry business will begin to be realized in the market.

Terafab Collaboration: Building Semiconductor Infrastructure with Elon Musk

Chen Liwu revealed that the Terafab project, promoted with Elon Musk, stems from a shared judgment—semiconductor infrastructure development lags behind AI demand growth in capacity, efficiency, and power consumption. Under this cooperation, Musk decided to build his own wafer fab, and Intel will provide technology and process support to help accelerate production. Chen Liwu said he holds weekly meetings with Musk’s team, and the collaboration is progressing smoothly.

He also mentioned that Musk has unconventional ideas at the operational level, such as discussing whether smoking should be allowed in certain cleanroom areas. "I might not go that far, but some areas perhaps could be, the key is to keep an open mind," he said.

Market Misconception: Intel Still in the "Crawling" Stage, True Potential Post-2030

In response to market skepticism about Intel’s transformation progress, Chen Liwu cited his consistent "crawl-walk-run" framework. He said that in recent months, Intel is still in the "crawl" stage: quietly building teams for CPU architecture, GPU architecture, and software architecture, aiming for leapfrog innovation at the speed of a "big startup"; in foundry, the gap with TSMC remains significant, and humility is necessary to strengthen IP, yield, and other foundational capabilities.

"My VC intuition tells me—look for 10x returns," Chen Liwu said. He referenced his experience at Cadence: from acting CEO to retirement, creating about 76 to 85 times returns for shareholders. He admits that Intel’s larger scale makes it harder to replicate, but "achieving 10x returns within 5 to 10 years" is a clear goal he sets for himself.

Below is the transcript of the interview:

Host: Welcome back to No Priors. Today, Allad and I have invited Chen Liwu—legendary investor from Walden, former CEO of Cadence, now CEO of Intel. We’ll discuss his plans to transform Intel, what it means that the U.S. government has become a major shareholder, how to become an outstanding semiconductor investor, and whether we can manufacture chips within the U.S. Welcome, Liwu.

Why take on the burden at Intel?

Host: Let’s start with an obvious question. Taking on the CEO role at this extremely important American semiconductor company is a truly tough job. Why did you accept?

Chen Liwu: That’s a good question. I’m 66 this year. Many people say I should retire and not take on the industry’s hottest job. There are several reasons: first, this is an iconic company, extremely important to the entire semiconductor ecosystem and the U.S.; second, after Cadence, I decided to do one more big thing.

Host: A lot has happened over the past year. What surprised you the most?

Chen Liwu: The most surprising thing is something I’ve never experienced in any previous job or training—one morning, President Trump asked me to resign, citing conflicts of interest, with no exceptions. I first convinced myself: I don’t need this job; I’m doing this purely to save Intel. Putting personal emotions aside, I started thinking about what I could do for Intel. Fortunately, I secured a meeting on Thursday morning, another on Monday, and after explaining my background—born in Malaysia, raised in Singapore, graduated from MIT, and having lived in the U.S. ever since—I shared this with him. He listened, and gave me the chance to continue. I am very grateful.

Host: You said this job is "saving Intel." What does a thriving, successful Intel look like in your mind?

Chen Liwu: It’s been over 14 months, and a lot has happened. First, changing the culture, establishing clear accountability, speeding up decision-making. I’m used to startup pace—everything moving at light speed—but Intel has layers of bureaucratic meetings, which I had to change. Second, listening to customers—truly satisfying them requires humility, listening, facing their problems, and solving them. Third, from day one, I decided all engineering teams report directly to me. I’m an engineer by background—I need to know where problems are, what needs fixing. Listening to customers, making them happy, ensuring we have the right products, simplifying product lines, and setting a clear roadmap and vision for the next five to ten years.

Intel’s Ten-Year Vision

Host: What is your vision for Intel in ten years?

Chen Liwu: My usual approach—whether at Cadence or Intel—is to first crawl, stay humble, listen to customers; then walk; then run. Step by step.

The first step is to strengthen the balance sheet—honestly, it was in pretty bad shape at the time. I’m glad the U.S. government became a major shareholder. I explained to President Trump: look at Japan, Singapore—this is infrastructure, the government should support.

Second, I’m very grateful to my old friend Jensen Huang—he invested $5 billion in Intel, and I’m happy I did some valuable work. His $5 billion has now appreciated to $25 billion or more. Also, SoftBank’s Masayoshi Son—who I served on the board of—has lent a hand. Through these, we stabilized the balance sheet.

Next, focus on products—simplify lines, listen to customers, launch next-gen leading products. Fortunately, AI inference demand is booming now, so I caught a good timing. In the past, CPU-to-GPU ratio was about 1:8; now I see it shifting to 1:4 or even lower. CPUs are becoming more important, which makes me very pleased.

I’ve talked with some AI model developers—they say that in reinforcement learning and coordinating all agents, CPUs perform better. So now, CPU demand is very high for me. After laying a solid foundation in data center server products, another key business is our foundry. It’s capital-intensive and not easy. You need the right IP mix—for mobile customers, low-power IP; without these, you can’t serve them. It’s a service business, a trust business—if yield isn’t up to standard, customers will abandon you due to revenue loss. So I focus heavily on yield, defect density, cycle time, ensuring high quality and reliability. Ultimately, we aim for a full-stack approach—not just silicon, but software, too. Customers ask me directly, "Give me the whole rack," so system-level solutions are needed. I’m quietly pushing these forward, recruiting the best talent I can find. By the way, all hiring is personally done by me, without headhunters.

Collaboration with Elon Musk on Terafab

Host: Another widely discussed initiative is Terafab and the collaboration with Elon Musk. Can you tell us how this came about and how you work together?

Chen Liwu: I think we both agree that Elon Musk is one of the greatest entrepreneurs of this century. We share a view: semiconductor infrastructure development has lagged behind AI growth—in capacity, efficiency, and power consumption. Under this framework, Musk decided to build his own wafer fab, and Intel is happy to support him with technology and process expertise to accelerate production. I meet weekly with his team, and the progress is very good.

He also has unconventional ideas—like discussing whether smoking should be allowed in certain cleanroom areas. "I might not go that far, but some areas perhaps could be, the key is to stay open-minded," I said.

Global Semiconductor Supply Chain Changes

Host: From a macro perspective, how is AI driving changes in the global semiconductor supply chain? What are your observations?

Chen Liwu: AI’s impact on the landscape will surpass the internet, and its influence will be deeper. AI first makes tasks more efficient—thanks to many intelligent agents, many tedious tasks previously done manually can now be faster. For example, in semiconductor design, timing optimization and time-to-market can be greatly improved, reducing costs.

Growth in AI demand faces several bottlenecks: first, power limitations—some countries simply lack enough electricity; second, the impact of helium—many don’t realize helium’s significant effect on the industry; third, memory shortages—this is the most urgent issue—expanding capacity now takes years to come online, and CPU and GPU shortages drive prices up, passing costs to clients.

The most impacted companies are those that do not embrace AI. AI can help enterprises improve efficiency across almost all functions. Companies should proactively adopt AI, find better ways to utilize it—whether in prediction, design, or workload management.

Host: The simplest argument against Terafab and Intel’s foundry competitiveness is labor costs and the feasibility of domestic manufacturing. You’ve decided to continue investing in foundry—what’s the logic behind that?

Chen Liwu: When I decided to continue betting on foundry or to exit, there were many voices—saying it’s too expensive, it’s infeasible. But I judged that this is critically important for the U.S. and the entire industry.

We’ve all experienced supply chain challenges. No large semiconductor company can rely heavily on just one or two regions. Increasingly, people realize that manufacturing in the U.S. is vital.

Our most advanced process, like 18A—around 1.4 nm—we’re already planning for 1 nm and 0.7 nm nodes. As process nodes shrink, line widths are thinner than hair, complexity is extremely high, and any mistake can wipe out previous efforts. That’s why manufacturing precision requirements are rising, becoming a bottleneck.

We respect TSMC greatly; we are good partners. The industry needs more capacity to serve customers, so we’re determined to persevere—this is key for the long term and where I can create more value for the industry.

Physical Limits and Advanced Packaging

Host: Many have discussed that chip scaling will hit physical limits, with line widths becoming too narrow to continue shrinking. When do you think we’ll hit a real wall?

Chen Liwu: We have 18A, are pushing for 14A mass production, and see paths toward 10 nm and 7 nm. This is feasible, but it will become more expensive and difficult. That’s why we need partners—close collaboration with substrate suppliers and equipment vendors—to push yield and performance.

Another bottleneck is advanced packaging. TSMC has CoWoS; we have EMIB, a next-generation solution. I must ensure it reaches customer-required yields at mass production.

When traditional scaling hits a wall, I turn back to materials for breakthroughs—GaN, SiC, InP. I’ve invested in all three. In packaging materials, I focus on glass—excellent for heat dissipation and insulation. I invested in a company called 3DGS. Intel holds about 1,000 patents in modules; integrating substrates and modules is a key challenge. We recently announced cooperation projects in India and New Mexico for advanced packaging. I also pay attention to synthetic diamonds—another excellent insulating material—and invested in a diamond wafer company.

The engineer’s spirit is like this—constantly encountering bottlenecks, then finding ways to cross or bypass them. Having participated deeply in the entire semiconductor lifecycle—from EDA tools to design to manufacturing—I’m now glad to leverage this experience to contribute to the industry.

Host: Could process node convergence flatten performance differences among foundries, approaching an asymptote?

Chen Liwu: Moore’s Law is about transistor density doubling, but power and cost don’t necessarily decrease proportionally—performance can double, but area and cost may not. Unless new materials and design methods are found, this is why I’ve increased recruitment of materials science talent—this is where innovation in the field is centered.

Eighteen years ago, when I invested in semiconductors, many top VCs weren’t interested. I remember at partner meetings, after talking about semiconductors, half would leave, the other half asked, "Do you have software or service projects?" Only one or two would stay and sympathize. Now, Nvidia’s market cap is $5.3 trillion, Broadcom and TSMC are each $2 trillion, AMD is close to $800 billion, and Intel about $600 billion. Semiconductors have become a hot, indispensable foundation again. 15–20 years ago, almost no VC wanted to invest in semiconductors—except giants like Samsung, ARM, SoftBank. Now, VC enthusiasm is high, and I’m very pleased.

Challenges in Semiconductor Investment

Host: As a long-term investor and operator, you see many difficulties—capital intensity, unpredictable results, deep workload understanding, high switching risk for customers, cyclical industry… How do you view these risks, and what advice would you give others about where to invest in this supply chain?

Chen Liwu: Venture capital and startups are in my blood—I really enjoy it. Not to boast, but for context: I have 159 IPOs, 126 mergers and acquisitions, over 200 semiconductor investments, with 38% in the U.S.

My investment approach always starts with a core question: where is the bottleneck, what problem am I solving? For example, I invested in Cradle Semiconductor because interconnect became a bottleneck; in Celestial AI because optical interconnects are increasingly important in clusters—Jensen Huang invested in almost all photonics-related companies, which is no coincidence.

In design, I ask whether AI and machine learning can help reduce complexity and improve quality—I see huge opportunities in EDA. Several startups are working in this area, a gold mine. In new materials, GaN, SiC, InP are my focus; some have been acquired by ADI and others. Power management—converting from 40V to 1V involves huge losses—is another key bottleneck I watch.

My investment framework: does the problem truly exist? Are customers struggling with it? And most importantly: who is the first target customer? I prefer very large customers—they have the ability and willingness. If they like your product, they’re willing to pay millions or provide guarantees over the next few years. Securing a big customer allows scaling.

Talent is also critical—U.S., Silicon Valley, Austin, Israel are key focus areas. Israel has highly disruptive, innovative entrepreneurs, very diligent. Even in wartime, they keep meetings—sometimes saying "Alarm, I need to go to the basement, network might be bad, let’s switch to voice"—that resilience impresses me.

Besides AI, physical AI is the next frontier. Full-stack development is essential. That’s why I remain deeply involved in frontier model investments—I see great potential in open-source cutting-edge tech for physical AI, another gold mine.

Cadence Experience

Host: You mentioned AI enabling faster, cheaper, more creative chip design and testing. Based on your Cadence experience, which directions are most promising? Are some already showing results?

Chen Liwu: I spent nearly 15 years at Cadence. One of my proudest achievements is finding my successor—training him myself. He’s now an excellent CEO, actively embracing AI, integrating intelligent agents into tools to boost efficiency. Synopsys’ Sassine is doing the same, supported by Nvidia’s $2 billion investment, and has acquired Ansys to expand into full-system design.

Big companies are doing this, but startups also have opportunities to do more disruptive work. Ultimately, these startups can IPO or be acquired by giants. It depends on the founders’ vision. My philosophy: if they want quick exits, I help them; if they want IPO from day one, I help them pursue that. As VCs, we support founders’ dreams and help realize them.

Scaling and Investment Decisions

Host: The directions you mentioned—materials companies, EDA, manufacturing—if you look 10 years ahead, will Intel or future semiconductor companies be unrecognizable because of AI?

Chen Liwu: I believe so. Returning to your point about capital intensity, unpredictability, and cyclicality—these must be factored into investment decisions. I prefer early entry, building teams; finding investors who will stay with you through tough times, not just fair weather; and seeking strategic investors—those adding value in manufacturing, storage, interconnects, etc. I also have friends in growth and hedge funds with unique perspectives on public markets, helping startups avoid pitfalls.

Honestly, looking back, 9 out of 10 companies I invested in changed their business plans midway because the market shifted. I prefer teams that are open-minded, willing to listen, and accept advice, but ultimately form their own judgment—best outcome is not "he does what I tell him," but "he gives feedback, then deduces conclusions I agree with or understand." That’s the joy of entrepreneurship.

In 10 years, winners will be those focusing on niche areas, finding the right partners, and scaling effectively. Full-stack solutions are crucial. Big companies like Jensen Huang’s focus on CUDA and platform, building a platform company—he succeeded. Startups like Anthropic, OpenAI, can also change the game elegantly, moving at lightning speed to become dominant.

For Intel, I hope it plays such a role—integrating XPU, advanced packaging, and foundry to create custom chips for different workloads—that’s my vision.

AI Era Team Restructuring

Host: The software industry is changing rapidly—who to hire, managing multiple agents. Many now prefer hiring 30-50-year-olds, experienced in team management, which can transfer to managing AI agents. How do you see team structure and skills evolving in hardware or foundry backgrounds?

Chen Liwu: Returning to the crawl-walk-run framework. During the "crawl" phase, I recruited the best talent in semiconductors; now I’m thinking about what software talent to bring in for full-stack capabilities; also, the average team age is over 40 or 50—I need to introduce younger talent to understand workloads and cutting-edge open-source models.

Interestingly, my son has become my teacher. Visiting him and playing with my grandson, I ask him about AI and machine learning—he knows more than I do. I learn a lot, then try to incorporate these insights into investment and talent decisions.

Intel used to be a very traditional company relying on spreadsheets. I’m transforming it into an AI-enabled enterprise—not just in design, but organization-wide, embracing AI to reduce dependence on spreadsheets. We combine senior technical talent with AI tools, not only in sales and marketing but also in design.

Industry Policy and Capital Sources

Host: For capital-intensive companies, raising funds has always been a challenge. Industry policies have created giants like TSMC, but this approach has long been unpopular in U.S. business culture. What’s your view?

Chen Liwu: For capital-intensive businesses and infrastructure projects, access to capital is crucial. Now some VCs are willing to invest $1 billion in a single company—unthinkable before. Early-stage, you want to enter when valuations are reasonable; Series A is already over $1 billion, which is tough.

Capitals that can help scale—like mutual funds—are less sensitive to shareholding ratios, and I welcome such investors. For capital-heavy projects like AI factories and fabs, government funding, sovereign wealth funds, or large infrastructure funds are essential. Sovereign funds and government support will become increasingly important.

As a public company, I focus on long-term growth investors, not just short-term funds asking "When will you buy back stock?"—which is reasonable, but I must also build the business. Balancing these is key.

Biggest Investor Misunderstanding of Intel

Host: What do you think is the biggest misconception investors have about Intel right now?

Chen Liwu: Several points. First, returning to crawl-walk-run: in recent months, I’ve still been crawling, but people are starting to see the potential. In products, we still hold PC market share, but performance must be greatly improved—so I’m quietly building teams for CPU, GPU, and software architectures, preparing for leapfrog innovation, like a big startup, using better technology to leap ahead.

In foundry, we still lag behind TSMC; we must stay humble, focus on foundational work—IP, yield, defect density, cycle time—to make foundry more efficient and reliable. It’s a trust business—customers must trust you before handing over wafers. These things take time, but I believe by 2030–2032, people will see how big Intel’s true potential is.

PC client is our core, but we’re extending into edge, physical AI, and agent AI. Previously, we only served servers and PCs for humans, but now a whole new dimension emerges—millions of agents need computing power and software stacks. I believe Intel has opportunities in both physical AI and agent AI. The game isn’t over.

AI is just the beginning. You have Huang Renxun leading training, edge computing, agents AI, and physical AI—huge opportunities. Everyone still has a chance. The past 14 months have generated 6x returns for shareholders, but this is just the start—there’s much more room.

My VC intuition: look for 10x returns. At Cadence, from acting CEO to retirement, the stock rose from $2.40 to about 76x return for shareholders; as executive chairman, about 85x. Intel’s scale is bigger and harder to replicate, but my goal is 10x—achieve 10x returns within 5 to 10 years. As a VC-inclined person, that’s my target.

Where Will Computing Power Reside?

Host: Some believe data centers will keep growing, gigawatt scale is just the beginning, and centralization is mainstream. But your business outlook also includes edge and client computing. Do you think computing power will ultimately be distributed among data centers, edge, and clients, or will it depend entirely on workloads?

Chen Liwu: The current large-scale AI infrastructure buildout is correct—I see no reason to slow down because workloads are still growing. The main constraint is supply-side—any slowdown comes from supply chain constraints, not demand.

But I care more about: after all this infrastructure is built, what applications will run on it? You need to find truly scalable applications—like Amazon, Netflix in the internet era—some applications emerge victorious, others disappear or are acquired. The AI industry will go through a similar process: after rapid growth, there will be consolidation, and a few real winners will surface.

Focusing on applications is key. Netflix is a real application, Amazon is a real application—both succeeded. Some applications are better suited for edge or client: robots, defense scenarios, device-side computing power is critical. Your assumptions about connectivity and device capabilities determine what you can do. This was somewhat overlooked during the SaaS boom.

My investment approach: find real problems, find the right partners, assess if the market size is sustainable—if you truly believe, double or triple down. This also includes betting on applications not yet widely deployed.

Host: Thank you very much for today’s insights—this has been a real pleasure.

Chen Liwu: Thank you for having me.