TechFlowPost, citing Wallstreetcn, reported that Intel CEO Lip-Bu Tan used his first podcast interview to lay out a sweeping plan for rebuilding Intel. Tan said his target is to deliver a “10x” return over “5 to 10 years,” adding that Intel has already generated roughly a 6x return for shareholders over the past 14 months, while stressing that “this is only the beginning.” His roadmap is not limited to Intel’s traditional PC client base. It extends into data-center servers, edge computing, physical AI, agentic AI, and the integration of XPU, advanced packaging and foundry capabilities to provide customized silicon for different workloads.
Tan explained why he accepted the role at age 66. He said Intel is an iconic company that matters to the entire semiconductor ecosystem and to the United States, and that after Cadence he wanted to do one more major thing. He also described the most surprising episode of his tenure: President Donald Trump had asked him to resign early one morning over alleged conflicts of interest. Tan said he first put personal emotion aside and reminded himself that he did not need the job and was doing it to save Intel. After securing meetings, he told Trump that he was born in Malaysia, grew up in Singapore, graduated from MIT and had lived in the United States ever since. He said Trump listened and gave him the chance to continue.
Repairing the Balance Sheet and Refocusing Products
Tan repeatedly described Intel’s transformation through his “crawl, walk, run” framework. The first stage is to remain humble, listen to customers and move step by step. He said his early work at Intel included changing the culture, clarifying accountability and accelerating decision-making. Coming from a startup-oriented pace, he found Intel’s layers of meetings and bureaucracy to be something that had to change. Because he is an engineer by training, he decided from day one that all engineering teams should report directly to him, so that he could know where problems were and what needed to be corrected.
The balance sheet was one of the first issues he addressed. Tan said Intel’s balance sheet had been in poor condition, and he welcomed the U.S. government becoming a major shareholder. He said he explained to Trump that this was an infrastructure issue, pointing to Japan and Singapore as examples of government support. He also thanked Nvidia CEO Jensen Huang, saying Huang invested $5 billion in Intel and that investment has now grown to $25 billion or more. SoftBank’s Masayoshi Son, whose board Tan had served on, also helped. With that financial base strengthened, Intel moved to simplify its product lines, focus on customers and prepare next-generation leading products.
On product demand, Tan said agentic AI and inference are creating very strong CPU demand. In the training era, the CPU-to-GPU ratio was roughly one to eight. He now sees it moving toward one to four, and even lower. Conversations with AI model developers led him to the view that CPUs perform better in reinforcement learning stages and in coordinating and scheduling large numbers of agents. That has made Intel’s data-center server line an important foundation before the company expands further into system-level offerings, including software and full-rack solutions requested by customers.
Foundry Discipline, Trust and the Terafab Collaboration
Tan acknowledged that Intel’s foundry business is capital-intensive and difficult. Many outside voices argued that it was too expensive or unworkable, but he decided to continue. His reasoning is that advanced domestic manufacturing in the United States has strategic value for supply-chain security, and that large semiconductor companies cannot rely entirely on one or two geographically concentrated suppliers. Operationally, he has focused on yield, defect density and cycle time. Foundry, in his words, is a service business and a trust business: “Customers must trust you before they hand you their wafers.” If yields fall short, customers can lose revenue and leave.
Tan also said Intel and TSMC are partners rather than purely competitors, and that the industry needs more capacity to serve growing demand. Intel’s most advanced process includes 18A, described in the interview as a 1.4-nanometer-class technology, and the company is planning for 1 nanometer and 0.7 nanometer. As nodes shrink, line widths become extremely fine and process complexity rises sharply. Tan said any error in the manufacturing flow can undermine the whole effort, making manufacturing precision an increasingly central bottleneck.
The Terafab collaboration with Elon Musk came from a shared judgment: semiconductor infrastructure has not kept pace with AI growth in capacity, production efficiency or power efficiency. Musk decided to build his own fab, and Intel will provide technology and process support to help accelerate production. Tan said he meets with Musk’s team every week and described the collaboration as energizing. He also noted Musk’s willingness to challenge convention, including discussions about whether smoking could be allowed in some cleanroom areas. Tan said he would not go that far, but that maintaining an open mind and evaluating unconventional ideas matters.
Advanced Packaging, Glass Substrates and Synthetic Diamond
As traditional node scaling moves closer to physical limits, Tan is turning to materials science and advanced packaging. Intel has 18A, is pushing 14A into production, and can see paths to 10 nanometers and 7 nanometers, he said, but that path becomes more expensive and more difficult. Advanced packaging is becoming a major constraint. TSMC has CoWoS, while Intel is pushing its next-generation EMIB approach. Tan said he must ensure EMIB reaches the yield customers require at production scale.
When conventional scaling hits bottlenecks, Tan said he goes back to materials. He has invested in gallium nitride, silicon carbide and indium phosphide. Some of the companies in those areas have already been acquired by large semiconductor firms such as ADI. In packaging materials, he has focused on glass. He invested in 3DGS because glass has valuable properties as a heat-dissipating insulating material. Intel also holds about 1,000 patents in modules, and integrating substrates with modules is one of the engineering issues Tan emphasized. Intel has recently announced advanced-packaging manufacturing cooperation projects in India and New Mexico in the United States.
Tan is also watching synthetic diamond. He invested in a synthetic diamond wafer company because diamond is another strong insulating material for chip packaging. He summarized the engineering mindset this way: “You keep encountering bottlenecks, and then you find a way to cross them or go around them.” He added that Moore’s Law is fundamentally about doubling transistor density, while power and cost do not automatically fall at the same rate. New materials and new design methods are therefore central, and he is increasing hiring in materials science for that reason.
How Tan Invests Across the Semiconductor Stack
Tan also discussed his record as a long-term investor and operator. He said he has been involved in 159 IPOs and 126 M&A exits, with more than 200 semiconductor investments, 38% of them in the United States. His investment framework starts with a simple question: where is the bottleneck, and what problem is being solved? He invested in Cradle Semiconductor because interconnect became a bottleneck. He invested in Celestial AI because optical interconnect is becoming more important inside clusters. He noted that Jensen Huang has invested in nearly every photonics-related company, saying that this is not a coincidence.
He sees major opportunity in EDA, where AI and machine learning can reduce complexity and improve design quality. He also highlighted power management, especially the losses involved in converting from 40V to 1V. For customer selection, Tan prefers a clear first target customer, often a hyperscale customer that has the ability and willingness to pay millions over several years or provide some form of commitment if the product is compelling. He also emphasized talent in the United States, Silicon Valley, Austin and Israel, praising Israeli entrepreneurs who continue holding meetings during wartime, even when they have to move to shelters and switch from video to audio.
His Cadence experience remains an important reference point. Tan spent nearly 15 years at Cadence and said one of his proudest achievements was selecting and training his successor, who is now embracing AI and bringing agentic AI into tools to improve efficiency. Synopsys’ Sassine is doing similar work, with $2 billion of investment support from Nvidia, and Synopsys has acquired Ansys to expand toward full-system design. Tan said large companies are moving, but startups still have room to do more disruptive work and eventually go public or be acquired, depending on the founder’s ambition.
AI-Native Teams, Capital Sources and Long-Term Investors
Inside Intel, Tan is also rethinking teams for the AI era. In the “crawl” stage, he recruited top semiconductor talent. He is now considering what kind of software talent is needed to build full-stack capabilities. He noted that the average age of some teams is in the 40s to 50s, and he wants younger people who understand workloads and frontier open-source models. He also said his son has become his teacher on AI and machine learning. Intel, in his description, was an old-style company heavily dependent on spreadsheets; he wants to turn it into an AI-enabled organization across design and the broader company.
On capital, Tan said access to funding is crucial for capital-intensive businesses and infrastructure projects. Some venture capital firms are now willing to invest $1 billion in a single company, something he said was unimaginable before. Early-stage investors either need to enter very early at reasonable valuations or invest by the Series A stage, but he noted that some Series A valuations have already exceeded $1 billion. For AI factories and foundries, he said government funding, sovereign wealth funds and large infrastructure funds are necessary sources of support. As a public-company CEO, he is also looking for long-term growth-oriented investors, not only short-term investors focused on quarterly buybacks.
Intel’s 2030–2032 Window and Where Compute Will Live
Asked about the biggest investor misunderstanding of Intel, Tan returned to “crawl, walk, run.” He said Intel is still crawling. In PC clients, the company still has market share but must improve performance significantly. He is quietly building CPU architecture, GPU architecture and software architecture teams to move faster, like a large startup. In foundry, Intel remains far behind TSMC and has to stay humble while strengthening IP, yield, defect density and cycle time. He expects Intel’s real potential to become clearer around 2030 to 2032.
Tan sees Intel’s opportunity extending from the PC base into edge computing, physical AI and agentic AI. In the past, servers and PCs primarily served humans. Now millions of agents will need access to compute and software stacks. He compared his return target with Cadence: from acting CEO to retirement, Cadence’s share price rose from $2.40 and produced roughly a 76x shareholder return; by the end of his term as executive chairman, the return was around 85x. Intel is larger and harder to move, he said, but as someone whose instincts are rooted in venture capital, his target is still 10x over 5 to 10 years.
On where compute will reside, Tan said large-scale AI infrastructure buildouts are justified by continuing workload growth, with constraints now coming mainly from supply rather than demand. After the infrastructure is built, the key question is which applications will run on it at scale. He compared the process with the internet era, when Amazon and Netflix emerged as real applications while others disappeared or were acquired. Some workloads, including robotics and defense, fit better at the edge or on client devices, where assumptions about connectivity and onboard capability determine what can be built. For Intel, integrating XPU, advanced packaging and foundry capabilities to deliver specialized chips for different workloads is the long-term direction he described.