Intel CEO Lip-Bu Tan's Latest Interview: After Missing Mobile, Cloud, and AI, We Cannot Miss the Next Wave

Video Title: Intel CEO Lip-Bu Tan on the Comeback of American Chip Industry
Video Author: TechSurge: Deep Tech VC Podcast
Compiled by: Peggy, BlockBeats

Editor’s Note: In the context of generative AI driving a new wave of computing power investment, industry discussions are shifting from "who has the most powerful chips" to "who can organize a more complete computing system." As the demand for GPUs, advanced processes, and capital expenditures in data centers have become consensus, a more fundamental question begins to emerge: what determines the efficiency of the next stage of AI infrastructure— the performance of a single chip, or the synergy between computing, storage, interconnect, packaging, and manufacturing?

In a podcast under Celesta Capital, Tech Surge, host Michael Marks engages in a conversation with Intel CEO Lip-Bu Tan. The discussion extends from his early semiconductor investment experiences to Cadence's transformation, as well as Intel's product, manufacturing, and platform strategies in the AI era.

In this conversation, Lip-Bu Tan does not simply provide a product roadmap for Intel's recovery; rather, he dissects semiconductor competition into a set of more fundamental structural issues: why hardware has re-emerged as a focal point for capital, how AI bottlenecks spill over from chips to infrastructure, whether vertical integration can recreate system value, and how a large company that has missed multiple rounds of technological migration can restore its perception of frontier changes.

First, the value of semiconductors is returning from a single product to a technological foundation. Over the past 20 years, software has received higher valuations and more venture capital, while semiconductors were once viewed as unsuitable for venture investment due to long R&D cycles, heavy capital investment, and limited exit paths. Now, AI has turned computing power, power consumption, and bandwidth into direct constraints for application expansion. Chips are no longer just carriers for software operation but are the infrastructure that determines model costs and commercialization boundaries. This means that the return of hardware is not merely a valuation rotation but a re-pricing of capital driven by changes in technological bottlenecks.

Second, AI competition is expanding from single accelerators to system engineering. The last round of AI investment mainly revolved around GPUs and model training, but as workloads shift towards inference, agents, and physical AI, bottlenecks begin to spill over to CPUs, storage, high-speed interconnects, advanced packaging, and cooling. The shift from air cooling to liquid cooling and even microfluidic cooling, from electrical interconnects to photonic technology, and from traditional packaging to glass substrates and new materials reflects the same change: cluster efficiency can no longer be solved solely by improving the performance of a single chip. This also means that the next round of hardware value may not be concentrated solely in leading GPU companies but may be distributed across the weak links of the entire computing system.

Third, the key to Intel's insistence on vertical integration lies not in how many businesses it retains but in whether it can reorganize dispersed capabilities into a platform. In the past, the division of design and manufacturing drove the rise of the fabless model and specialized foundries. Intel's simultaneous operation of products and wafer foundry has increased the complexity of organizational and capital allocation. Lip-Bu Tan emphasizes that the combination of CPU, GPU, software, advanced packaging, and wafer foundry is crucial because product optimization in the AI era increasingly relies on cross-layer collaboration. Its potential value lies not only in self-manufacturing but also in jointly optimizing architecture, packaging, and processes based on customer workloads; the corresponding risk is that if product competitiveness and manufacturing execution cannot improve simultaneously, vertical integration may continue to amplify costs.

Fourth, the relationship between CPUs and storage is being redefined by AI. In the past, CPUs were Intel's most stable core business, while storage was often viewed as a cyclical product with significant price fluctuations. As AI shifts from training to inference, the demand for general computing has not disappeared; CPUs still need to handle data processing, task scheduling, and agent operation. Meanwhile, memory bandwidth, capacity, and power consumption have become important constraints on system performance. Lip-Bu Tan mentions CPU and storage stacking and new storage architectures, indicating that the focus is not necessarily on Intel returning to the traditional storage market but rather on the need for computing and memory to re-synchronize at the architectural and packaging levels.

Fifth, what Intel truly needs to fix may not be a specific generation of products but its ability to acquire external information and respond quickly. The management approach established by Lip-Bu Tan at Cadence was to directly listen to employees and customers, transforming the relationship between the company and customers from "suppliers" to "partners" that can share roadmaps. For Intel, reconnecting with customers, universities, AI laboratories, venture capital firms, and startups is also about restoring this perceptual capability. A large tech company missing a technological wave often does not mean it completely fails to see new directions, but rather that external changes cannot be timely translated into internal resource allocation and product decisions.

If we were to compress this conversation into a single judgment, it would be: the competition in AI hardware has shifted from a race of single-point performance to a system competition determined by computing, storage, interconnect, packaging, manufacturing, and organizational capabilities. In this sense, the subject of this discussion is no longer just whether Intel can complete a corporate transformation, but whether a traditional chip giant can re-establish its ability to participate in the next generation of computing platforms.

The following is the original content (for ease of reading and understanding, the original content has been reorganized):

TL;DR

Lip-Bu Tan believes that semiconductors have re-emerged as the core of the tech industry, and AI competition has expanded from a single chip to encompass the entire stack of packaging, storage, interconnect, cooling, and software.

He brings the transformation experience from Cadence to Intel: stay humble, listen to customers, respond quickly, and upgrade customer relationships from "suppliers" to "partners."

Intel will continue to retain the vertical integration model of product design, advanced packaging, and wafer foundry to create greater value for customers through platform capabilities.

CPUs remain at the core of Intel's rebuilding competitiveness, with agent AI, inference, edge computing, and physical AI potentially driving a new wave of demand.

Lip-Bu Tan reveals that Intel is researching CPU and storage stacking and new storage architectures but is not yet ready to announce specific plans.

After missing the waves of mobile internet, cloud computing, and AI, Intel will reconnect with universities, venture capital firms, and startups to avoid falling behind in frontier innovation again.

Key Points from the Interview

Global semiconductor sales are approaching $1 trillion ahead of schedule.

In a conversation on the Tech Surge podcast under Celesta Capital, Intel CEO Lip-Bu Tan stated that AI is making hardware re-emerge as the core of the tech industry. However, this opportunity is no longer limited to GPUs but extends to CPUs, storage, advanced packaging, high-speed interconnects, photonic technology, and cooling systems.

For Intel, this is not just a product cycle but a reconstruction of platform capabilities.

Lip-Bu Tan candidly admits that Intel has previously missed significant waves such as mobile internet, cloud computing, and AI. Therefore, his goal now is: "From now on, I will not miss any major wave again."

Semiconductors are returning to the center of AI competition from being a "sunset industry".

Lip-Bu Tan began investing in chips in 1987, accumulating investments in nearly 550 companies. However, for a long time, semiconductors were not a favored direction for venture capital firms.

He recalls that 20 years ago, when visiting top venture capital firms, the entire partner team would often be present at the start of meetings. But when he began discussing semiconductors, half of the people would politely find excuses to leave, leaving only a few "sympathetic" listeners.

At the end of the meeting, they would usually ask him, "Do you have any software or service startups?"

At that time, mainstream venture capital firms viewed semiconductors as a sunset industry, and capital continuously shifted towards software and internet services. Even some of Lip-Bu Tan's investors believed that his continued investment in chips while others exited was a near-crazy choice.

However, Lip-Bu Tan believes that chips are always the underlying foundation of the tech industry. Without chips and platforms that are suitable in terms of performance, power consumption, and cost, many upper-layer applications cannot exist.

This judgment has also made him persist as a contrarian investor.

He mentions that in the past, a co-investor once asked him, "Can you name a semiconductor company with a market value of over $1 trillion?" Today, that question is no longer valid, as semiconductor companies have emerged among the most valuable tech companies globally.

But Lip-Bu Tan is not only focused on large companies like Nvidia. In his view, semiconductors represent a vast technological system, and many key innovations come from lesser-known small companies: some are reducing Chiplet power consumption, some are solving high-speed interconnect issues, and others are betting on photonic technology, advanced packaging, and new cooling materials.

The real investment opportunities often lie within these bottlenecks that have not been fully recognized.

Investing in SambaNova: AI Inference Cannot Rely Solely on High-Power GPUs

Lip-Bu Tan's judgment on AI hardware is a reflection of his investment approach.

Having invested early in graphics chip companies like S3, he recognized the high power consumption issue of GPUs early on. He also judged that as AI shifts from model training to actual deployment, the market size for inference and agent AI could far exceed that of training.

About nine to ten years ago, he supported two different computing architecture paths for this reason.

The first was Cerebras's wafer-scale chip solution. Lip-Bu Tan believed that this technology was highly challenging to implement, but the problem the founder Andrew Feldman was trying to solve was worth supporting, so he invested from the A round.

The second was SambaNova's RDU, or Reconfigurable Dataflow Unit. Its dataflow architecture attempts to provide another path for AI computing outside of GPUs while maintaining computational performance and reducing power consumption.

In 2017, under Lip-Bu Tan's push, Celesta made its first investment of $2 million in SambaNova, when the company's valuation was about $12 million. Since then, he has participated in multiple rounds of financing for SambaNova and helped the company attract new investors.

Lip-Bu Tan stated that SambaNova is advancing its Series F financing, with the amount expected to be between $800 million and $1 billion. This refers to the scale of financing, not the company's valuation.

He emphasized that investing in startups should not be solely about betting on a single founder, but rather about seeking a complete team that can continuously adjust its direction. Markets change, and among the ten companies he invested in, about nine will alter their initial business plans during their development.

What truly deserves long-term support is a team that can adapt to changes, establish the right culture, and ultimately build world-class enterprises.

Transforming Cadence: Turning Suppliers into Collaborative Partners

When Chen Liwu took over Cadence, the company's stock price had already fallen to about $2.42.

He initially agreed to serve as the interim CEO for just three months, while the company simultaneously searched for a permanent leader. However, those three months eventually turned into 15 years. During this time, Cadence completed a transformation of its corporate culture and product strategy, and its stock price significantly rose from its low point.

Chen Liwu summarized the core of this experience in three words: humility, listening, and responsiveness.

When he first took on the role of CEO, he told employees at a company-wide meeting, "This is my first time as a CEO, and if you have any good ideas, please send them to me." Subsequently, he received about 300 emails daily and replied to each one. For suggestions that warranted deeper understanding, he would even walk over to employees' desks to communicate directly.

This approach helped him identify information silos within the company and allowed management to hear real feedback from the product frontlines.

Customer relationships also needed to change. Chen Liwu recalled that some of Cadence's customers were very angry at the time, demanding refunds and explicitly stating they were unwilling to continue using the company's products. Other customers complained that their product issues had gone unanswered in the past, with the company team only appearing when contract renewals were approaching.

As a result, Chen Liwu pushed Cadence to establish a rapid response mechanism. Later, one customer told him that less than 24 hours after filing a complaint, someone had already come to the office to resolve the issue.

A major competitor of Cadence once told him, "The same customer sees me as a supplier but views you as a partner."

In Chen Liwu's view, this is the key distinction between the two types of relationships. Once customers see a company as a partner, they are more likely to share product roadmaps and real needs. Companies can also combine feedback from other customers to provide more valuable suggestions.

He is bringing this approach to Intel, although Intel's business is more complex: the company needs to rebuild product competitiveness while also driving the success of its foundry business.

Intel Needs to Build Chips and Reconstruct Computing Platforms

When asked why Intel needs to simultaneously handle chip design, manufacturing, and sales, Chen Liwu's answer was: the combination of products, advanced packaging, and foundry services can create greater value for customers.

In theory, Intel could further shift to an outsourcing model and stop manufacturing chips in-house. However, Chen Liwu still believes in vertical integration because future computing competition is not just about comparing individual chips, but about the synergy of entire systems.

However, this route is only viable if Intel first has sufficiently competitive products.

Chen Liwu acknowledged that Intel once held a very strong market position in the CPU and computing fields, but has made many mistakes over the past years, gradually losing some advantages. Rebuilding Intel requires attracting excellent CPU architects, GPU architects, system architects, and software talent to establish full-stack capabilities from chips to systems and then to software.

The CPU remains the core of this strategy.

As AI moves from training to inference and further into intelligent AI, the demand for general-purpose CPUs may increase again. Chen Liwu stated that he frequently receives calls from other company CEOs who hope Intel can provide more CPUs. On one hand, Intel needs to increase supply; on the other hand, it also needs to develop new CPU architectures to meet future workload requirements.

This wave of demand comes not only from traditional servers and data centers but will also extend to PCs, edge computing, and physical AI. Intel also needs to connect more closely with leading research institutions, AI labs, and software development ecosystems to drive chip and system design based on actual application needs.

AI Bottlenecks Overflow: Interconnect, Packaging, and Cooling Need to Be Redone

Chen Liwu believes that as the scale of AI computing expands, bottlenecks are shifting from the chips themselves to the surrounding infrastructure.

First is high-speed interconnect. As AI cluster scales grow, the performance of a single chip can no longer determine overall efficiency; data transmission between chips, servers, and racks becomes increasingly important.

Based on this judgment, he has invested in companies like Credo Semiconductor and Astera Labs and has also laid out plans in the field of photonic interconnects. Some related companies were later acquired by Marvell, Credo, and others.

Second is cooling. As the power consumption of CPUs and other AI chips continues to rise, cooling methods are shifting from air cooling to liquid cooling and further extending to microfluidic cooling.

Advanced packaging follows the same logic. Intel already possesses packaging technologies like EMIB-T, and Chen Liwu is also focusing on new materials such as glass substrates and synthetic diamonds, hoping to improve the packaging, insulation, and cooling capabilities of high-performance chips.

These technologies do not necessarily need to become independent businesses within Intel.

Chen Liwu's thinking is that technologies that can be developed internally should be advanced by Intel itself; those that are not suitable for internal development can first support the growth of external startups, and in the future, be integrated into Intel's platform through cooperation, integration, or acquisition.

What Intel needs to establish is not a set of scattered products, but a larger platform covering computing, interconnect, packaging, and manufacturing.

Returning to Storage? Intel is Exploring CPU and Memory Collaboration

Intel was founded in 1968 and initially engaged in the storage chip business, not microprocessors.

When asked whether Intel might re-enter the storage market, Chen Liwu did not disclose a clear plan but sent a signal worth noting: Intel is researching CPU and memory stacking, as well as new storage architectures.

In the past, he was not fond of investing in storage chips because traditional storage products have strong cyclical and commodity characteristics. However, as AI computing presents new demands for bandwidth, capacity, power consumption, and packaging, storage is transitioning from standardized components to a key component of system performance.

Chen Liwu stated that new technologies are making the storage industry different. Exploring new storage architectures has become one of the projects he is focusing on.

He also mentioned that he has hired former SK Hynix CEO Lee Seok-hee. As for whether this personnel arrangement means Intel will re-expand its storage business, he stated that the company is not yet ready to announce specific plans.

However, from an overall strategic perspective, Intel is not simply thinking about returning to the traditional storage market, but whether CPU, storage, packaging, and manufacturing can form a new system architecture.

Missing out on Mobile, Cloud, and AI, Intel Bets on the Next 15 Years

Chen Liwu's takeover of Intel is not merely a career choice.

He said that at his age, he could have easily retired. However, Intel is both an iconic company and holds significant importance for the semiconductor industry and the United States, so he hopes to participate personally and make a real impact.

This also determines his time scale.

When he joined Intel, he told the board that he is not someone who only focuses on the short term. He is considering how Intel can build a larger platform in 10 or 15 years and how this platform can truly benefit the entire industry.

This long-term perspective also influences his understanding of competitive relationships.

Chen Liwu has known industry figures like Micron CEO Sanjay Mehrotra and NVIDIA CEO Jensen Huang for many years. Today, there is both competition and investment and cooperation among these companies. NVIDIA has already become an investor in Intel, and the U.S. government and SoftBank have also entered Intel's shareholder list.

However, Chen Liwu stated that he does not simply view these individuals as competitors. The market is large enough, and the more important question is how to jointly create a larger market.

For Intel, the real challenge is not to temporarily lead in a particular round of product competition, but to re-enter the forefront of technology where new advancements are continuously emerging.

This means connecting university professors, startups, venture capital firms, and AI labs to find the next generation of opportunities from new materials, computing architectures, and system bottlenecks.

As Chen Liwu said, Intel has already missed major waves such as mobile internet, cloud computing, and AI. Now, he hopes to ensure one thing: when the next wave arrives, Intel must be present.