Written by: Techub News Compilation

Introduction

In May 2025, Sundar Pichai, CEO of Alphabet and Google, rarely accepted an in-depth interview with the All-In Podcast. During the hour-long conversation, Pichai addressed the central concerns in the current market regarding the “disruptor’s dilemma” that Google may face amid the AI wave. As the leader who has stewarded this tech giant for nearly a decade, Pichai's statements not only serve as a concentrated response to external doubts but also systematically explained how Google plans to harness its more than twenty years of technological accumulation—from infrastructure to cutting-edge research—to navigate and lead this AI revolution. The dialogue covered a wide range of topics including the future of search, AI business models, chip strategies, energy constraints, quantum computing, and robotics, providing a key window for understanding Google's strategic direction in the next decade.

Summary

• Pichai firmly believes that AI is an “extraordinary opportunity” for search, not a zero-sum game disruptor; new features such as AI Overviews are expanding the types and use cases of search queries.
• Google's core advantage lies in its full-stack technological accumulation, from TPU chips and a global data center network to cross-disciplinary foundational research, enabling it to promote model innovation while controlling costs.
• In the face of intense competition from OpenAI, Microsoft, Meta, and others, Pichai believes that the market created by AI will be larger than the sum of all previous technologies, with continued innovation and execution being the key to success.
• Google is making long-term, patient investments in areas like quantum computing and robotics (physical AI), viewing them as foundational breakthroughs akin to early AI, and expects significant advancements in the next 3-5 years.
• Pichai acknowledges that cultural management is a persistent challenge, with recent emphasis on regaining “mission focus” to ignite the intensity and optimism similar to Google's early startup phase to cope with competition in the AI era.

AI is the Evolution of Search, Not Disruption

The interview opened with a direct question: is Google’s core search business facing a real risk of disruption in an era where conversational AI is reshaping human-computer interaction? Sundar Pichai’s answer was clear and resolute. He recounted that since becoming CEO in 2015, he has positioned Google as an “AI-first” company because the team believes that AI will be the driving force behind the greatest advancements in search. Thus, for him, the current moment is not a threat but rather an “extraordinary opportunity” for search.

Pichai substantiated his viewpoint with data. The “AI Overviews” function, launched by Google about a year ago, is now used by over 1.5 billion users across more than 150 countries. This has not only not cannibalized traditional search but has actually expanded the types and scope of user queries, bringing about entirely new use cases. In queries that trigger AI Overviews, Google has observed a growth in query volume that is persistent. He revealed that a completely new dedicated search AI experience called “AI Mode” is being tested in the lab and will be detailed at the Google I/O conference. In this mode, users will be able to conduct complete conversational follow-up questions, with Google's most advanced models using search as a native tool to answer queries.

“Users are responding with their actions,” Pichai pointed out, with the average length of user queries now two to three times that of two or three years ago, sometimes even long paragraphs. This is changing the nature of search. Regarding the classic proposition of the “innovator's dilemma,” Pichai argued that the dilemma only exists when you perceive it as such. He cited the example of mobile transformation and how YouTube continued to thrive after the rise of TikTok, noting that every major innovation cycle in tech history requires full commitment. “It’s a bit like one of Google's original principles: follow the user; everything else will follow.”

Transitioning Business Model: Controllable Costs, Advertising Logic Remains

When the conversation shifted to the sensitive topic of the business model—whether more expensive AI queries would disrupt the delicate unit economics between “query costs” and “per-query ad revenue” in Google Search—Pichai exhibited confidence based on infrastructure advantages.

He acknowledged that two years ago, there were widespread concerns about costs, but he has always believed that, in terms of service costs, he is willing to bet that Google's infrastructure capabilities can outperform almost all other companies. The facts have proven this, as the cost of servicing a single query has significantly decreased over the past 18 months. Currently, the main constraint is latency, not costs. How to achieve search speeds that are nearly instantaneous in AI experiences is the more important frontier issue at present.

Regarding ad revenue from AI queries, Pichai stated that in the AI Overviews where ads have been displayed, their performance benchmarks are now on par with traditional search. “We have reached that stage.” He further analyzed from first principles: ads work effectively in search because commercial information is itself part of user intent. There is no reason why AI cannot perform better in this area. Therefore, he feels “reassured” about a smooth transition of the business model, noting that while some adjustments may take time, all indicators show that Google can successfully complete this transition.

Infrastructure and Chips: Google’s “Full-Stack” Moat

Pichai repeatedly emphasized that one of Google’s core and often underestimated advantages is its two decades of infrastructure investment. From early shippable container data centers to the global network in place today, this forms the cornerstone of competitiveness. In the AI era, this advantage is directly reflected in the Pareto frontier of performance and cost—Google can offer the best models at the most cost-effective prices.

One of the key components is Google's self-developed TPU chips. Pichai mentioned that Google launched its first-generation TPU in 2017, and it has now developed to the seventh generation. The latest “Ironwood” chip has a computational power exceeding 40 Exaflops. These chips, specifically optimized for machine learning, perform exceptionally well, particularly in inference, which is a significant reason why Google is confident in serving AI search queries at a massive scale. He stated that Google's “full-stack” approach—from low-level infrastructure and foundational R&D to upper-layer application innovation—will provide long-term advantages to the company.

On chip strategy, Pichai clarified the relationship with NVIDIA. He praised NVIDIA as an “extraordinary company,” and described CEO Jensen Huang as outstanding, noting their long history of collaboration. Google internally uses TPUs to train Gemini models, serves part of its traffic on GPUs, and provides options for cloud customers. He believes that companies in the industry will try similar (self-developed chip) paths, but NVIDIA's research capabilities and software stack are world-class, possessing numerous advantages. Google's long-term strategy is to maintain the flexibility of using both TPUs and GPUs, allowing them to mutually promote and push the technology frontier.

Model Innovation Not at a Bottleneck, Data Advantage Lies in Personalization

In response to external claims that the performance of large language models is beginning to “plateau” and that competitors are narrowing the differences, Pichai disagrees. He compares the progress process to “artificial jag jag intelligence,” indicating that advancement is not always linear and may experience periods that seem slow, followed by paradigm breakthroughs. In recent years, the industry has continuously advanced in areas such as pre-training, post-training, and inference computing, and is now exploring how to integrate all technologies into intelligent workflows. He feels that progress is quite continuous.

“Progress will indeed become more difficult, which will distinguish elite teams,” Pichai believes, asserting that the more difficult the problems, the better prepared Google is. Google’s research frontier is far broader than most companies, not limited to LLM or transformer-based models, and deep explorations are taking place across various fields, including diffusion models. He has not yet observed a fundamental turning point where the return on computational investment begins to diminish.

When asked if Google has a unique data advantage from products like YouTube, Pichai's response focused on user experience rather than purely on data training. He believes Google’s opportunity lies in leveraging the personal context generated when users use Gmail, Calendar, Docs, YouTube, Search, and other products (with user permission) to provide better quality, more personalized experiences. He views this as one of the differentiating innovation opportunities for the company’s future, which is currently being cautiously advanced.

Future Interaction: Seamless, Ambient Computing and AR Glasses

Looking ahead to human-computer interaction over the next 5 to 10 years, Pichai envisioned a “seamless” future where the hard adaptations that humans need to make will increasingly wane, and computation will actively serve humans. He believes this is the “Holy Grail,” and technologies such as touch and voice are gradually moving towards this future.

He is particularly optimistic about the prospects of augmented reality glasses. As a glasses wearer himself, he has personally experienced AR glasses, which, while currently less comfortable than regular glasses, are improving. He believes that when AR truly matures, it will elevate seamless experiences to new heights, allowing computation to exist “ambiently” and serve users. He is not referring to immersive display devices but rather to AR glasses. Combining truly multimodal models, he finds this paradigm very interesting and capable of evoking a sense of the next leap. He compares the current stage of AR development to that of smartphones around 2006-2007, suggesting that it might still require several cycles to reach its ideal state.

Pichai confirmed that Google has invested a significant amount of time in hardware, not limited to Pixel phones and data centers, but also feels excited about new forms like AR glasses and robotics.

Competitive Landscape: The AI Market is Big Enough, Founders Are Deeply Involved

Faced with a powerful matrix of competitors comprised of OpenAI (Sam Altman), xAI (Elon Musk), Meta (Mark Zuckerberg), and Microsoft (Satya Nadella), Pichai exhibited an open and pragmatic attitude. He praised this as an “impressive group” that brings together the best companies and entrepreneurs. The existence of competition suggests that we will witness significant advancements.

He particularly mentioned a meeting with Elon Musk in May 2025, admiring his “unparalleled willpower to turn future technologies into reality.” Pichai emphasized that AI will open up a larger opportunity landscape than the sum of all past technologies. He reminded people that Google didn’t even exist when the internet emerged. Future winners in the AI field may include many companies that have yet to emerge. Therefore, the landscape is not simply about “who defeats whom,” but rather about which companies can succeed in partnership with the best talent through innovation and execution.

Pichai also shared the deep involvement of Google co-founders Larry Page and Sergey Brin. He frequently communicates with both, praising their early foresight into today's AI and considering it one of the most exciting moments in computer science. Sergey Brin has even been “hardcore” in his involvement, coding alongside the Gemini team, reviewing loss curves, and providing feedback on model architecture, infusing unmatched energy into the team. Pichai values the inspiration that comes from the “non-linear thinking” collision with the two founders.

Long-term Bets: Quantum Computing and Robotics

Google is known for its long-term, patient investments in cutting-edge technologies such as TPU, DeepMind, and Waymo. Pichai views quantum computing and robotics through the same lens.

He compares the current development stage of quantum computing to AI around 2015. He predicts that within the next five years, there will be a true “epiphany moment”—completing a highly useful practical computation in a quantum manner that far surpasses classical computers, thus truly showcasing the potential of this field. He is very confident in this. Google’s goal is to showcase more practical algorithms through its cloud platform and provide access. Pichai believes that just as the field of self-driving has been noisy in the past, the quantum field currently has a lot of announcements, but Google considers itself to be at the forefront.

In the field of robotics, Pichai admitted that Google had attempted to apply layers too early (such as acquiring Boston Dynamics and then selling it) at a time when robotics had not yet been deeply affected by AI. Now, the convergence of “AI + Robotics” has arrived. Google has a world-class cutting-edge research team in robotics, and Gemini's efforts in vision-language-action models are also top-notch. They are considering how to participate through collaboration or product launches. Pichai observed that the progress of humanoid robots is so rapid that he needs to carefully discern the authenticity of videos. He predicts that the “magical moment” in the field of robotics may be two to three years away. Google’s subsidiary Intrinsic is committed to becoming the “Android of the robotics field,” and the Gemini model will also become a powerful foundational model in robotics.

Culture, Talent, and Future Outlook

In response to external criticism that Google’s culture has evolved from an early innovation paradise to a “spoiled employees” refuge, Pichai defended and reflected on this. He explained that the intention behind providing benefits such as lunch is to create a positive, optimistic, and cross-disciplinary innovation atmosphere, which remains a source of strength for Google to attract top talent and inspire grassroots innovation. However, he also acknowledged that as the company has expanded rapidly, sometimes secondary issues overshadow the company’s core mission.

In recent years, he has intentionally focused on regaining “mission focus,” emphasizing that employees gather to innovate and serve the company's mission, making an impact. The high intensity and excitement brought about by the current AI moment remind him of early Google. He has created mechanisms such as “labs” to encourage small teams to innovate rapidly. He believes that culture requires constant fine-tuning to stay true to core values and must be actively realigned when drifting occurs.

Regarding talent acquisition, Pichai indicated that Google remains competitive in attracting top PhD researchers, with returning employees as well. He anticipates that AI will lead to the emergence of exceptional talent from more regions around the globe, not just a few traditional centers.

On the structure of Alphabet as a holding company, Pichai clarified that its essence is not merely financial investment but aims to solve problems across different fields based on common foundational technologies (like AI), giving rise to seemingly disparate yet interconnected businesses. From Waymo to Isomorphic (the drug discovery company based on AlphaFold), this logic is exemplified.

Finally, when asked about his greatest pride and regret during his decade-long tenure as CEO, Pichai expressed immense pride in Google’s ability to push the technological frontier (such as achieving Nobel-level foundational research) while translating it into commercial value. As for regrets, he mentioned several acquisitions that were intensely debated but ultimately uncompleted, such as Netflix, but he prefers to look forward, considering these as “another possibility in a parallel universe.”

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。