Intel CEO Lip-Bu Tan used his first podcast interview to present a broad restructuring plan for the company, framing his ambition in venture-capital terms: a 10x return over five to ten years. Tan, the Walden veteran, former Cadence CEO and current head of Intel, said the company is rebuilding its technology roadmap around advanced packaging, new semiconductor materials, next-generation substrate technology and foundry execution as conventional process-node scaling moves closer to physical limits.
Tan said Intel has already generated roughly a sixfold return for shareholders over the past 14 months, but added that “this is only the beginning.” He pointed to 2030 to 2032 as the period when outsiders will begin to see Intel’s full potential. In his view, the company’s future should not be confined to its traditional PC client base. Intel also has to move into edge computing, physical AI and agentic AI, while combining XPU, advanced packaging and foundry capabilities to produce custom chip solutions for different workloads.
Why Tan Took on Intel
Asked why he accepted the role at age 66, Tan gave two reasons. Intel is an iconic company that remains important to the semiconductor ecosystem and to the United States, and after Cadence he wanted to do one more major thing. He described the job as an effort to “save Intel.” His early priorities were cultural change, clear accountability and faster decision-making. Intel’s layered meeting structure and bureaucratic processes, he said, had to be replaced by a pace closer to that of a startup.
Tan also recalled an unexpected episode in which President Donald Trump asked him to resign early one morning, citing a conflict of interest. Tan said he first put aside his personal emotions and focused on what he could do for Intel. He later secured a meeting and explained to Trump that he was born in Malaysia, grew up in Singapore, graduated from MIT and had lived in the United States ever since. Tan said Trump listened and gave him the opportunity to continue, for which he expressed gratitude.
His management style has been deliberately engineering-centered. From the first day, Tan decided that all engineering teams would report directly to him. As an engineer by background, he said he wanted to know personally where problems were occurring and what needed to be corrected. He summarized Intel’s ten-year path with a “crawl, walk, run” framework: first be humble, listen to customers and stabilize the balance sheet; then simplify the product line and deliver leading next-generation products; finally connect CPU, GPU, software architecture, foundry and system-level capabilities into a broader platform.
Balance Sheet, CPU Demand and the Foundry Trust Problem
Tan said Intel’s balance sheet was in poor condition when he arrived. He welcomed the U.S. government becoming a major shareholder and compared that kind of support to the way Japan and Singapore support infrastructure. He also thanked Nvidia CEO Jensen Huang, who invested $5 billion in Intel; Tan said that investment has since grown to $25 billion or more. SoftBank’s Masayoshi Son, whose board Tan had previously served on, also provided support. These moves, according to Tan, helped stabilize Intel’s financial base.
On the product side, Tan said agentic AI and inference workloads are creating strong renewed demand for CPUs. In earlier training environments, the ratio of CPUs to GPUs was roughly one to eight; he now sees that ratio moving toward one to four, or even lower. After speaking with AI model developers, Tan said he heard that CPUs can perform better in reinforcement learning and in coordinating and scheduling large numbers of agents. That has made the CPU relevant again in data center servers, while Intel continues to build out its foundry business and system-level solutions.
Tan repeatedly described foundry as both a service business and a trust business. “Before customers hand you their wafers, they have to trust you,” he said. He identified yield, defect density and cycle time as the key indicators. If yields fail to meet expectations, customers suffer revenue losses and leave, and winning them back becomes extremely difficult. Although critics question labor costs and the feasibility of domestic manufacturing, Tan said he chose to keep investing in foundry because no major semiconductor company should rely too heavily on one or two geographically concentrated supply sources. In his view, advanced manufacturing in the United States has strategic value for the industry and for supply-chain security.
Terafab, Musk and Global Supply Bottlenecks
Tan also discussed Intel’s Terafab work with Elon Musk. The project, he said, grew out of a shared view that semiconductor infrastructure has not kept pace with AI growth in capacity, production efficiency or power efficiency. Musk decided to build his own wafer fab, and Intel is providing technology and process support to help accelerate production. Tan said he meets Musk’s team every week and described the collaboration as energizing. Musk, he noted, constantly asks why things must be done in the traditional way. Tan also mentioned one unconventional discussion about whether smoking could be allowed in certain areas of a cleanroom. Tan said he would not go that far, but the important point is to keep an open mind and evaluate ideas carefully.
From a global supply-chain perspective, Tan said AI will affect the world more deeply than the internet. He identified several bottlenecks: power constraints in some countries, the impact of helium on the semiconductor industry, and the current shortage of memory. Even if new capacity is added now, it takes years to come online. CPUs and GPUs are also in short supply, pushing prices higher and eventually passing costs to clients. Tan said the companies hit hardest will be those that do not embrace AI, because AI can improve efficiency across forecasting, design and many other workloads.
Physical Limits, EMIB and New Materials
On the limits of process scaling, Tan said Intel already has 18A and is pushing 14A into mass production. He can see a path to 10 nanometers and 7 nanometers, but he said that path will become more expensive and more difficult. That is why advanced packaging has become a central focus. TSMC has CoWoS, while Intel is promoting its next-generation EMIB approach, and Tan said he must ensure it achieves customer-required yields in mass production. Intel has also announced advanced-packaging manufacturing cooperation projects in India and in New Mexico in the United States.
When traditional miniaturization runs into bottlenecks, Tan said he returns to materials science. He has invested in gallium nitride, silicon carbide and indium phosphide, and some of those companies have been acquired by large semiconductor companies such as ADI. In packaging materials, he has focused on glass because it is a strong thermal insulating material, and he invested in 3DGS. Intel has roughly 1,000 patents in modules, and integrating substrates with modules is one of the engineering problems he emphasized.
Tan also invested in a synthetic diamond wafer company, seeing diamond as another strong insulating material for chip packaging. “That is the spirit of engineers — you keep running into bottlenecks, then you find a way to cross them or go around them,” he said. He added that Moore’s Law is fundamentally about doubling transistor density, but power and cost do not fall at the same rate unless new materials or new design methods are found. That is why he is recruiting more materials-science talent.
Investment Discipline, Cadence Lessons and Intel’s Misunderstanding
As a long-time investor, Tan cited his own record: 159 IPOs, 126 M&A exits and more than 200 semiconductor investments, with 38% of them in the United States. His investment framework begins with one question: where is the bottleneck, and what real problem is being solved? He invested in Cradle Semiconductor because interconnect became a bottleneck, and in Celestial AI because optical interconnect is becoming more important inside clusters. He also sees major opportunity in EDA, where AI and machine learning can reduce design complexity and improve design quality.
Tan repeatedly referred to his Cadence experience. He spent nearly 15 years there and said one of his proudest achievements was identifying and training his successor, who is now bringing agentic AI into design tools to raise efficiency. Synopsys, led by Sassine, is moving in a similar direction, supported by a $2 billion investment from Nvidia and the acquisition of Ansys as it expands into full-system design. For startups, Tan said his philosophy as a VC is to support the founder’s dream: if founders want a quick exit, help them achieve it; if they want an IPO from day one, help them pursue that path.
On team structure, Tan said Intel is still in the “crawl” phase. He has recruited top semiconductor talent and is now considering what kind of software talent is needed to build full-stack capability. He also noted that the team’s average age is in the 40s to 50s, and he wants to bring in younger people who understand workloads and frontier open-source models. Tan said Intel has historically been an old-school, spreadsheet-dependent company, and he is trying to turn it into an AI-enabled enterprise across the whole organization, not only in design but also in sales, marketing and operations.
Tan also addressed capital. For capital-intensive businesses and infrastructure projects, he said access to funding is critical. AI factories and foundries need support from government funds, sovereign wealth funds or large infrastructure funds. As a public company CEO, he said he is deliberately looking for long-term growth-oriented investors rather than only short-term holders asking each quarter when the company will buy back stock. Shareholder returns matter, he said, but so does building the business.
Asked what investors misunderstand most about Intel, Tan returned to the “crawl, walk, run” model. Intel is still crawling, and the company remains far behind TSMC in foundry. It must stay humble and strengthen IP, yield, defect density and cycle time before customers fully trust it with wafers. Those efforts take time, but Tan believes that by 2030 to 2032 people will begin to see Intel’s real potential. The PC client business remains the base, but Intel is moving toward edge, physical AI and agentic AI. Tan said his VC instinct is to look for 10x opportunities. At Cadence, he said the stock moved from $2.40 to about a 76x shareholder return during his CEO tenure, and roughly 85x by the time he finished as executive chairman. Intel is larger and harder to transform, but his target remains a 10x return over five to ten years.
At the end of the interview, Tan discussed where compute will reside. He supports the current large-scale AI infrastructure buildout and sees no demand-side slowdown, saying any deceleration would come from supply constraints. But he is also focused on which applications will run on that infrastructure once it is built. Robotics and defense, for example, can be better suited to edge or client-side compute, where assumptions about connectivity and on-device capability shape what can be done. For Intel, Tan said the path is to combine XPU, advanced packaging and foundry capability to build specialized chips for different workloads.