NVIDIA: OpenAI, Future of Compute, and the American Dream | BG2 w/ Bill Gurley and Brad Gerstner

Jensen and Clark Tang discuss the transformative potential of OpenAI and Nvidia’s strategic partnership, projecting OpenAI as a future multi-trillion dollar company. They highlight Nvidia's role in AI infrastructure, emphasizing the integration of training and inference through reinforcement learning. The conversation covers Nvidia's annual release cycle, extreme co-design, and competitive advantages in AI infrastructure. They also discuss the geopolitical implications of AI development, stressing the importance of US-China relations and advocating for a balanced approach to global AI competition. Jensen underscores the need for the US to leverage its technological leadership while fostering global collaboration.

• OpenAI is predicted to become a multi-trillion dollar hyperscale company, similar to Meta and Google, by offering consumer and enterprise services.
• The opportunity to invest in OpenAI before it reaches this scale is considered a smart investment due to the company's potential for high returns.
• Nvidia's partnership with OpenAI involves building AI infrastructure and data centers, enabling OpenAI to function as a fully operational hyperscale company.

"I think that OpenAI is likely going to be the next multi-trillion dollar hyperscale company."

• OpenAI's growth trajectory positions it as a future leader in the hyperscale market.

"The opportunity to invest before they get there, this is some of the smartest investments we can possibly imagine."

• Investing in OpenAI now is seen as a strategic move due to its expected exponential growth.

• Three scaling laws are identified: pre-training, post-training, and inference.
• Post-training involves AI practicing skills until mastered, integrating training and inference through reinforcement learning.
• The new inference model emphasizes thoughtful response generation, improving answer quality through research and learning.

"We now have three scaling laws, not one."

• The evolution from a single scaling law to three enhances AI's capability and efficiency.

"The longer you think, the better the quality answer you get."

• Emphasizing thoughtful processing in AI leads to improved outcomes.

• Nvidia is partnering with OpenAI to build AI infrastructure, including data centers and system-level integrations.
• This partnership will support OpenAI's exponential growth in customers and computational demands.
• OpenAI's self-building of data centers aims for a direct relationship with Nvidia, similar to Elon Musk's approach with his companies.

"We're partnering with OpenAI to build their own self-build AI infrastructure for the first time."

• Nvidia's collaboration with OpenAI is crucial for developing OpenAI's self-sufficient AI infrastructure.

"They're going through two exponentials...the usage exponential and the computational exponential."

• OpenAI's growth is driven by increasing demand and computational needs, necessitating robust infrastructure development.

• General-purpose computing is shifting to accelerated computing and AI, marking the end of Moore's Law.
• This transition involves replacing traditional CPU-based systems with GPUs for enhanced performance.
• The shift is expected to refresh trillions of dollars of computing infrastructure globally.

"General-purpose computing is over, and the future is accelerated computing and AI computing."

• The industry is moving away from traditional computing paradigms toward more efficient, accelerated models.

"Our partnership with Intel is recognizing that general-purpose computing needs to be fused with accelerated computing."

• Collaborations with industry leaders like Intel are essential for transitioning to accelerated computing.

• AI is expected to augment human intelligence, potentially increasing global GDP by trillions of dollars.
• The integration of AI into various sectors will enhance productivity and efficiency, similar to past industrial revolutions.
• AI's role in generating tokens and augmenting human labor is likened to historical advancements like electricity and jet engines.

"AI supercomputers...are going to generate tokens to augment human intelligence."

• AI's capacity to enhance human productivity is significant and transformative.

"That $50 trillion is going to get augmented by something."

• The potential economic impact of AI is substantial, with significant growth anticipated in global GDP.

• AI-driven revenues are projected to grow from $100 billion to at least $1 trillion by 2030.
• The transition from CPUs to AI-driven systems is already underway in hyperscale companies.
• The growth of AI infrastructure and applications is expected to continue at a rapid pace.

"The probability that we're going to have a trillion dollars of AI revenues by 2030 is near certain."

• The forecast for AI revenue growth is optimistic, with substantial increases expected in the coming years.

"The hyperscalers, they went from CPUs to AI...their entire revenue base is all now AI-driven."

• Major companies have already shifted to AI-driven models, contributing to the industry's rapid expansion.

• Despite concerns of potential overspending, investments in AI are viewed as essential for future competitiveness.
• The demand for AI infrastructure continues to outpace supply, with companies consistently underestimating growth forecasts.
• Nvidia and its partners are prepared to scale production to meet increasing demand for AI technologies.

"We're at the end of the supply chain, right? And so, we respond to demand."

• Nvidia's strategy is to adapt to demand fluctuations, ensuring supply meets the industry's needs.

"Every one of their forecasts that they provide us turns out to have been wrong because they under forecasted."

• Companies have consistently underestimated the demand for AI, leading to ongoing adjustments in production and supply.

• The conversation highlights two exponential growth phases in AI: the initial adoption and engagement phase, and the subsequent reasoning phase.
• The transition from AI's ability to memorize and generalize to reasoning marks a significant leap in AI capabilities.
• The increased reasoning capabilities of AI have led to a substantial rise in compute requirements.

"We've already had one exponential we were experiencing which was the adoption rate of AI, the engagement of AI was growing exponentially. The second exponential that kicked in was reasoning."

• This quote emphasizes the two distinct phases of exponential growth in AI: initial adoption and the subsequent reasoning capability, which has significantly increased computational demands.

• There's a significant shift anticipated from CPU-based data processing to AI-driven data processing.
• AI data processing will require domain-specific acceleration layers and recipes.
• This transition represents a massive market opportunity for companies like Nvidia.

"In the future, that's all going to move to AI data. That is one gigantic massive market that we're going to move to."

• The quote highlights the anticipated shift from traditional CPU-based data processing to AI, underscoring the vast market potential of AI-driven data processing.

• Companies investing in AI are seeing exponential revenue growth, driven by increased confidence in sustained revenue streams.
• The investment side of AI companies is viewed as an opportunity, separate from revenue generation.

"Their revenue which is growing exponentially. It has to be funded by their capital, the money they've raised through equity and whatever debt they can raise."

• This quote explains the financial dynamics of AI companies, where exponential revenue growth is supported by capital raised through equity and debt, reflecting investor confidence.

• Nvidia's annual release cycle has been crucial in keeping up with the exponential growth in AI token generation.
• The cycle involves significant upgrades, requiring comprehensive system redesigns at both hardware and software levels.

"The token generation rate is going up exponentially. And the customer use is going up exponentially."

• The quote illustrates the rapid increase in AI token generation and usage, necessitating Nvidia's annual release cycle to maintain performance and cost-effectiveness.

• Extreme code design involves optimizing models, algorithms, systems, and chips simultaneously.
• Nvidia's approach to extreme code design has allowed it to achieve significant performance improvements beyond what Moore's Law would predict.

"Extreme code design means that you have to optimize the model, algorithm, system, and chip at the same time."

• This quote defines extreme code design, highlighting its comprehensive approach to optimizing various components of AI systems to achieve superior performance.

• The competitive landscape in AI is more challenging than ever, with increasing wafer costs and the need for extreme code design.
• Nvidia's focus has shifted from building GPUs to creating entire AI factories, encompassing a wide range of components and systems.

"You're building a chip that's a component of a much larger machine. You've built a very sophisticated system, platform, factory."

• The quote reflects Nvidia's strategy of developing comprehensive AI infrastructure, moving beyond individual components to entire systems and platforms.

• The debate between GPUs and ASICs is ongoing, with Nvidia emphasizing the complexity and scale of AI infrastructure that goes beyond individual chips.
• Nvidia's systems approach is seen as more adaptable to the rapidly evolving AI landscape compared to ASICs.

"Most of them, given the seeming success of Google's TPUs, how are you thinking about this evolving landscape today?"

• This quote sets the stage for discussing the evolving landscape of AI hardware, particularly the challenges and opportunities in choosing between GPUs and ASICs.

• Nvidia's approach to AI involves disaggregated workload orchestration, allowing for flexibility and collaboration with other companies like Intel.
• This strategy enables Nvidia to integrate various components into its AI ecosystem, enhancing performance and adaptability.

"We announced this thing called Dynamo, disaggregated AI workload orchestration and we open sourced it because the future AI factory is disaggregated."

• The quote highlights Nvidia's innovative approach to AI workload management, emphasizing the importance of flexibility and collaboration in future AI factories.

• Nvidia argues that even if competitors offer chips at zero cost, the total cost of operating a system with Nvidia's technology is more economically viable.
• The efficiency and performance of Nvidia's systems translate to higher revenue potential for customers.

"Because the land power and shell is already $15 billion, we've taken a crack at kind of the math on that."

• This quote underscores the economic rationale behind choosing Nvidia's systems, where operational efficiency outweighs the initial cost of competitor chips.

• The discussion emphasizes the importance of performance per watt in data centers, highlighting the high opportunity cost of using less efficient chips.
• The competitive advantage of certain chips, such as Black Wall, is emphasized due to their superior performance metrics.

"You would always choose the best perf per watt."

• The quote underscores the necessity of selecting chips that offer the best performance per watt to maximize efficiency and output in data centers.

• The conversation explores how moving to an annual cadence and co-developing with the supply chain increases competitive advantage.
• The scale of development and acquisitions, like Envy Fusion and CPX, contributes to a growing competitive moat.

"Our opportunity as I described it is much larger than the consensus."

• This quote reflects the belief that the company's growth potential is significantly underestimated by the market, suggesting a robust competitive position.

• NVIDIA is positioned as an AI infrastructure company rather than just a chip manufacturer, emphasizing partnerships like the one with OpenAI.
• The flexibility in purchasing options for customers is highlighted, allowing them to buy specific components as needed.

"NVIDIA is really an AI infrastructure company."

• This quote clarifies NVIDIA's broader role in the AI landscape, extending beyond chip manufacturing to encompass comprehensive AI infrastructure solutions.

• The necessity for countries to develop sovereign AI capabilities is discussed, with AI seen as crucial to national security and economic future.
• The importance of encoding cultural and historical values into AI is emphasized.

"Nobody needs atomic bombs. Everybody needs AI."

• This quote illustrates the critical role AI plays in modern society, likening its importance to that of historical advancements like nuclear technology.

• The discussion touches on the collaboration between industry and government, particularly in the context of AI development and national security.
• The role of government in supporting technology growth and the strategic importance of AI is highlighted.

"This is an administration that believes in growth."

• This quote reflects the administration's focus on economic and technological growth as essential components of national security and global leadership.

• The competitive dynamic between the US and China in AI is explored, with an emphasis on the need for the US to remain competitive.
• The potential consequences of restricting technology exports to China are discussed, highlighting the importance of maintaining access to the Chinese market.

"We have a competitive relationship with China."

• This quote acknowledges the complex and competitive nature of US-China relations in the technology sector, emphasizing the need for strategic engagement.

• The impact of the $100,000 fee per H1B visa on talent recruitment is considered, with the potential implications for both large and small companies.
• The unique brand reputation of America as a destination for realizing the "American dream" is highlighted.

"America has one a singular brand reputation that no country in the world has."

• This quote emphasizes America's unique position as a destination for global talent, underscoring the importance of policies that support talent recruitment.

• The concept of the "American Dream" is tied to the idea of starting from nothing and achieving success through hard work and opportunity.
• Legal immigration is viewed as a fundamental part of the American Dream, whereas illegal immigration presents challenges that need pragmatic solutions.
• The introduction of a high price tag on H-1B visas is seen as a way to curb abuse, though it may favor large companies over startups.

"We are utterly singular and you're talking to somebody who represents the American dream. My parents didn't have any money. Sent us over here. We started from nothing."

• This quote illustrates the personal narrative of achieving the American Dream through hard work and opportunity, highlighting its significance.

"President Trump knows that we want legal immigrants. There's a difference between legal immigrants and illegal immigrants."

• Emphasizes the distinction between legal and illegal immigration and the importance of maintaining legal immigration as part of the American Dream.

• The need for a strategic plan to recruit and retain the world's best talent is emphasized to maintain the American Dream.
• High costs for H-1B visas could deter startups and encourage investment outside the U.S.
• The importance of creating conditions for foreign students and professionals to stay in the U.S. is highlighted.

"Charging $100,000 or let's say, you know, it got lowered to 50 or whatever the case is, it does seem like it tilts the playing field in favor of big companies."

• Discusses the potential negative impact of high visa costs on startups and the broader implications for the U.S. economy.

"It might accelerate investment outside United States, right? And so there are unintended consequences."

• Highlights the risk of driving investment away from the U.S. due to restrictive immigration policies.

• The decline in interest from international students, particularly from China, poses a challenge to maintaining U.S. leadership in innovation.
• The need to create an environment that welcomes and retains top talent is critical for the future success of the U.S.

"Three years ago 90% of the top AI researchers graduating from universities in China wanted to come to the United States and did... today that's closer to 10 or 15%."

• Illustrates the significant drop in interest from Chinese researchers, indicating a potential threat to U.S. innovation.

"Smart people's desire to come to America and smart students' desire to stay, those are what I would call KPIs."

• Identifies the attraction and retention of talent as key performance indicators for future success.

• The U.S. should maintain confidence in its ability to compete globally without resorting to isolationist policies.
• Collaboration with other countries and companies, even former competitors, is seen as beneficial for mutual growth.

"Bring it on. Because I believe in our people. I believe in our culture. I believe in our country."

• Reflects confidence in the U.S.'s ability to compete and thrive on the global stage.

"We just did the ultimate deal, partnering with Intel, a company that spent most of its life trying to put us out of business."

• Demonstrates the potential for collaboration with former competitors to achieve greater success.

• The concept of "Invest America" aims to give every child a stake in the country's future prosperity.
• Reindustrialization and technological advancements are seen as key drivers of economic growth and opportunity.

"Every child born forever more in the history of this country will start off with an investment account at birth."

• Describes a new initiative to ensure all children benefit from the country's economic growth.

"Reindustrializing America is just fundamentally going to be transformative."

• Emphasizes the potential impact of bringing manufacturing and industrial jobs back to the U.S.

• AI is viewed as a great equalizer that can close the technology divide and create new opportunities.
• While AI may change or eliminate certain tasks, it is expected to create new jobs and drive economic growth.

"AI will change tasks. The thing that people confuse is there are many tasks that will be eliminated. There are many tasks that will actually be created."

• Highlights the dual nature of AI's impact on the workforce, with both challenges and opportunities.

"The ultimate equalizer. We've closed the technology divide. Now they just have to learn human."

• Suggests that AI can democratize access to technology and opportunities, leveling the playing field.

• The rapid pace of technological change requires adaptability and a willingness to embrace new opportunities.
• The U.S. benefits from having visionary leaders and a robust system of risk capital to drive innovation.

"If you have a train that's about to get faster and faster and go exponential, the only thing that you really need to do is get on it."

• Encourages embracing the accelerating pace of change and innovation rather than resisting it.

"It's an extraordinary time. But I also think, you know, one of the things that I'm just grateful is that we have leaders who also understand their responsibility."

• Acknowledges the importance of responsible leadership in navigating the challenges and opportunities of technological advancement.