“The low-hanging fruit is gone, […] the hill is steeper,” Sundar Pichai, CEO of Alphabet and Google, admitted during his appearance at The New York Times’ DealBook Summit on December 5.
In a wide-ranging discussion with Andrew Ross Sorkin, Pichai painted a picture of both promise and complexity as artificial intelligence (AI) evolves. While dismissing the idea of an AI plateau, he acknowledged that the pace of progress would slow, requiring deeper technical breakthroughs to address increasingly intricate challenges.
“I don’t fully subscribe to the wall notion,” Pichai stated, referring to the theory that AI development is nearing its limits. “I think the models are definitely going to get better at reasoning, completing a sequence of actions more reliably, you know, more agentic if you will […] but you’re definitely going to need deeper breakthroughs as we go to the next stage.”
A Full-Stack Commitment to AI
Pichai said that at the heart of Google’s AI strategy is a “deep full-stack approach” that integrates foundational research, infrastructure, and product applications. Pichai detailed how this philosophy underpins Google’s competitive edge, from its Tensor Processing Units (TPUs) to cutting-edge models like Gemini.
“We are in our sixth generation of tensor processing units,” Pichai explained. These custom chips are critical for training and running large-scale models, enabling Google to optimize both performance and cost. Designed to handle massive parallel computations, TPUs form the backbone of Google’s machine learning infrastructure, supporting applications across Search, YouTube, and Google Cloud.
Related: Nvidia B200 Blackwell Outshines Google Trillium TPU in Latest MLPerf Training Benchmark
Gemini AI, Google’s newest generative model, exemplifies this integration. Launched earlier this year, Gemini merges text, image, and multimodal capabilities, allowing users to interact with AI across multiple formats.
Pichai highlighted a recent breakthrough: “Just coming today, we just announced groundbreaking research on from a text and image prompt creating a 3D scene”, pointing to the recent release of DeepMind´s Genie 2 model that creates dynamic 3D simulations, changing in real time based on user input. “The frontier is moving pretty fast”, he said.
He pointed at Google´s technological foundation with a broad ecosystem of products. Google’s portfolio includes 15 services, each with over 500 million active users, ranging from Maps to Gmail. Developers are also a key audience, with over three million building applications using Google’s AI models.
Transforming Google Search with AI
Pichai described how AI is reshaping this cornerstone of the business, making it more dynamic and responsive to complex queries. Search remains Google’s flagship product, accounting for the majority of Alphabet’s revenue.
Pichai admitted that Google “applied AI aggressively to search”. Recent advancements include multimodal search capabilities, enabling users to combine text and images to refine results. New features like AI-generated overviews and Gemini-powered tools have enhanced the search experience for over a billion users worldwide, in his opinion.
Related: Google Reduces AI Overviews Presence in Search Results to Just 15%
Looking ahead, Pichai forecasted that “Search itself will continue to change profoundly in ’25”. “I think we are going to be able to tackle more complex questions than ever before. You know, I think you’ll be surprised even early in ’25 the kind of newer things search can do compared to where it is today,” he remarked, emphasizing that the platform would continue to evolve to meet users’ needs. In a world flooded with AI-generated content, Google aims to remain a trusted intermediary, filtering information to deliver accurate and reliable results.
Rising Competition in the AI Space
Despite Google’s dominance, competition in the AI industry has intensified. Pichai addressed pointed criticism from Microsoft CEO Satya Nadella, who recently claimed that Google should have been the “default winner” in AI given its resources and early start. “I would love to do a side by side comparison of Microsoft’s own models and our models any day, any time. They’re using someone else’s models.,” Pichai countered, referring to Microsoft’s partnership with OpenAI.
While current-generation large language models (LLMs) have achieved relative parity among leading players like OpenAI, Anthropic, and Meta, Pichai expressed confidence in Google’s ability to lead the next phase. The next generation of LLMs “will stand out in ’25, so I think it’s an exciting year from that perspective,” he said.
Related: ChatGPT’s o1 Pro Mode Falls Short in SimpleBench, Will GPT-4.5 Turn the Tide?
The competitive dynamics extend beyond models to include deployment and user adoption. Pichai pointed to Google’s ability to integrate AI across its ecosystem as a key differentiator, a capability he argued few rivals could match.
“We are building foundational models and we give it, use it internally, we provide it to developers – over 3 million developers. And you know, it’s a deep full stack investment. We are getting ready for our next generation of models. I just think there’s so much innovation ahead. We are committed to being at the state-of-the-art in this field and I think we are.”
Google’s Position in the AI Landscape
As competition intensifies in the generative AI market, Pichai acknowledged the growing prominence of rivals like OpenAI, Anthropic, and xAI. However, he emphasized Google’s unique position as a vertically integrated company with expertise spanning research, hardware, and product ecosystems.
When asked about emerging players, Pichai highlighted the dynamism of the industry. Pichai also praised Meta’s work on the Llama model and acknowledged xAI’s potential, given Elon Musk’s track record.
“I think xAI given Elon’s track record for sure. Meta has done a great job with Llama, right? […] I mean these are extraordinary companies, right? So you know, I have… I mean we partner with Apple in certain areas and you know, I expect both Apple and Amazon – just had an event yesterday where they announced their own model.
So you know, these are companies with deep capabilities and deep access to capital and extraordinary track record of execution, which shows how dynamic a moment is. But I think it’s a mistake to think for the most profound technology we are going to work on as humanity, there aren’t going to be a handful of companies working on it, right? The opportunity space is much bigger than most people can process it today.”
Related: xAI Expands Colossus Supercomputer Tenfold to 1 Million GPUs
Despite these challenges, Google’s ability to integrate AI across its vast portfolio gives it a competitive edge thinks Pichai. From the development of Tensor Processing Units (TPUs) to the deployment of multimodal models like Gemini, he sees Google uniquely positioned to deliver AI capabilities at scale.
Pichai emphasized the need for patience and persistence. While breakthroughs in reasoning and agentic capabilities are expected, achieving these goals will require significant investment in both compute power and algorithmic innovation. He acknowledged the role of external factors, such as the availability of high-quality training data and improvements in computational efficiency.
Navigating Ethical Challenges in AI Development
As AI becomes more embedded in everyday life, ethical considerations are taking center stage. Pichai stressed the importance of balancing innovation with responsibility, citing tools like SynthID as part of Google’s effort to ensure AI is developed and deployed ethically. SynthID embeds invisible watermarks in AI-generated images, helping to combat misinformation and deepfakes.
“We are working on – we have SynthID, we have watermarking, we all but you know, it’s a research area. Over time like how does society think about like deep fakes that cause real harm to individuals etc.? Like you know, should you have some notions? […] you have to be thoughtful and deliberate and and be specific about what you’re trying to do.“
Related: Google Expands SynthID to Include AI-Generated Text and Video
Google’s commitment to ethical AI also extends to licensing agreements with content creators. Addressing concerns about fair use in AI training, Pichai said, “we license data from Reddit, right? We license data from AP, we license data from New York Times, right? And so you know, we do it across the world in variety of ways,” adding that in the future, there could be marketplaces where creators are compensated for their contributions to AI models.
This acknowledgment points to a future where intellectual property and AI co-exist within a structured framework.
The broader ethical question of AI’s societal impact remained a focus for Pichai. Pichai highlighted AlphaFold’s role in predicting protein structures, enabling breakthroughs in drug discovery, cancer treatment, and vaccine development. He emphasized that this showcases AI’s potential to drive meaningful progress.
“I’m definitely on the optimistic side. I you know, when I think about the areas where this technology is going to make a difference, you know, for for people by now I think the other area where we won… Demis and John Jumper won the Nobel was for AlphaFold right?
Where we used AI to solve protein folding, predict protein folding better. I mean those things are being used by people to discover better drugs, right? Tackle problems like cancer and vaccines and so on. So there are many many areas in which AI is going… so I think technology is going to make progress and and and that arc of progress is bigger than any one person alone so you know, no one person does.“
Related: Google DeepMind Open-Sources AlphaFold 3 to Researchers
Antitrust and Regulatory Scrutiny
Beyond ethics, regulatory challenges loom large for Google. The U.S. Department of Justice (DOJ) is pursuing an antitrust case that could lead to the breakup of the company’s core operations, including its search engine and Chrome browser.
While Pichai refrained from commenting directly on ongoing litigation, while he defended Google’s track record.
“I don’t have any particular insights into that, right? You know, I think we’ll defend ourselves there. But I just viewed as when I think the decade ahead you know, AI ends up being the biggest opportunity ahead and so I’m focused on what we can do to innovate as a company and I think if we can do that well, I think our success is in our hands more than anything else.”
Related: DOJ Moves to Split Google From Chrome Amid Antitrust Findings
Critics argue that Google’s agreements with companies like Apple to make its search engine the default option on devices stifle competition.
Pichai countered that the digital landscape is more competitive than ever, with platforms like ChatGPT and TikTok offering alternative ways for users to access information. “Competition is a click away,” he remarked, emphasizing that Google operates in an environment where consumer choices drive success.
Related: New ChatGPT Live Search Feature Challenges Google with Real-Time Web Access
Internationally, Google faces additional scrutiny from regulators in Europe and Asia, where concerns about privacy and data sovereignty have prompted calls for stricter oversight.
Pichai acknowledged these challenges but argued that existing regulatory frameworks, such as those governing healthcare and finance, already provide robust safeguards. “So you really want to be careful about what additional regulation if anything you need at all, right?
Like you know, you have to get your drugs approved – there’s an established process to do that, right? So I would take… AI is going to be used for everything from recommending a coffee shop next to you to maybe deciding your insurance to a health decision, right? So I think you know, the phrase regulating AI can be overbroad,” he said.
Internal and External Criticism
Pichai also addressed criticism of Google’s corporate culture and its perceived caution in adopting aggressive strategies. Andrew Ross Sorkin pointed to remarks from the late Charlie Munger, who once described Google’s headquarters as “a very rich kindergarten.”
In response, Pichai highlighted achievements like Waymo’s expansion, which he argued reflect Google’s ability to execute on long-term goals.
“I just think you have to see the track record of what we have done as a company and where we are headed. You know, just this year alone, take something like Waymo – as you know, people used to question whether you can do this. We kind of were at 50,000 rides like 6 months ago, 100,000 autonomous rides per week, 150,000 last week, it was 175. We crossed a million autonomous rides!
You know, four weeks ago I took my dad in San Francisco and put him in the car. He was in the front, I was sitting in the back with my mom driving around San Francisco. I mean to him you would have described as what we think of as AGI. We kind of just crossed that in San Francisco – the Waymo cars drive better than humans. You don’t do these things by not executing well, right?”
Cultural shifts within Google have also been a focus for Pichai. Reflecting on a period of employee activism and internal dissent, he described efforts to realign the company’s workforce around its core mission.
Pichai remarked that workplaces are not platforms to resolve personal differences but are spaces where employees should focus on shared missions and creating impactful products and services.
This recalibration has coincided with a broader industry trend toward prioritizing productivity and innovation over activism. “We see people are the happiest when they feel they’re being productive and the work is being impactful and they’re doing well,” Pichai observed, suggesting that these factors now outweigh other considerations in shaping workplace culture.
The Economics of AI: Licensing and Monetization
A recurring theme in the discussion was the economics of AI, particularly how content creators and platforms can coexist in an AI-driven ecosystem. Pichai outlined Google’s approach to licensing, which includes agreements with news organizations and social media platforms. “Over time will there be models by which people can create… there’ll be a marketplace in the future,” he said.
This approach reflects broader industry trends. For instance, Google has deliberately limited the availability of its music-generation tools to professional artists, ensuring that creators retain control over how their work is used.
Related: OpenAI Considers Ads Model to Boost Revenue, Says CFO Sarah Friar
Pichai suggested that similar models could be applied across other domains, from text generation to video production. “In YouTube we put a lot of effort into understanding and you know, identifying content and with Content ID and and creating monetization for creators. I think those are important principles right?
I think there’s always going to be a balance between understanding what is fair use when a new technology comes versus how do you give value back proportionate to the value of the IP, the hard work people have put in.,” he noted.
Looking ahead, Pichai predicted that AI would unlock opportunities across sectors, particularly in fields requiring advanced reasoning and specialization. However, he cautioned that achieving this potential would depend on overcoming technical barriers and aligning innovation with societal values.
The Next Decade: A Vision for AI’s Evolution
Pichai also addressed concerns about the scalability of AI infrastructure. While computational power has traditionally been a limiting factor, he argued that the real breakthroughs will come from algorithmic and technical innovations.
Pichai explained that while computational power is limited by available capital, the true differentiation in AI development lies in achieving technical and algorithmic breakthroughs.
“When you start out quickly scaling up, you can throw more compute and you can make a lot of progress, but you’re definitely going to need deeper breakthroughs as we go to the next stage. So you can perceive it as there’s a wall or you perceive it as there’s some small barriers […] everyone is going to get more compute, right? You know, for compute you’re only constrained by money, right?
Like capital to do it. I think there’s a lot of capital around so I don’t think processing power… I mean, I think first of all the current amount of compute we’re using is just an arbitrary number. It’s not like we’re using a lot of compute – there’s no reason why it can’t keep scaling up, right?,” he explained.
Related: Amazon Challenges Nvidia’s AI Dominance with Ultracluster Supercomputer
According to Pichai, these advancements will require a balance between speed and responsibility. He reiterated his commitment to ensuring that AI is developed in a way that aligns with societal values. Pichai stressed that the primary challenge with AI is not its technical feasibility but ensuring its alignment with societal benefits.
Leadership in the Age of AI
Pichai’s leadership style, often described as thoughtful and deliberate, was also discussed during the summit. Reflecting on his journey as CEO, Pichai emphasized the importance of adaptability and resilience. “It’s what in 2015 I set the company to be AI first. It’s why I was excited to be a CEO to and I think the next decade is going to be the most interesting there,” he said.
In his opinion, this strategic focus has positioned Google as a leader in AI, even as it faces mounting competition and regulatory scrutiny. But he also acknowledged the complexities of managing a company at the intersection of technology and society.
The Road Ahead
As the conversation drew to a close, Pichai reflected on Google’s broader mission and its place in the evolving technological landscape. Pichai reiterated Google’s mission to organize the world’s information and described AI as the next major step in achieving that goal. Whether through advances in search, breakthroughs in healthcare, or partnerships with businesses, Google aims to lead with a vision that combines innovation with responsibility, he said.
While the path forward is filled with challenges, from competitive pressures to regulatory hurdles, Pichai remained optimistic about Google’s ability to navigate them.
