Thumbnail for Winning the AI Race Part 3: Jensen Huang, Lisa Su, James Litinsky, Chase Lochmiller

Winning the AI Race Part 3: Jensen Huang, Lisa Su, James Litinsky, Chase Lochmiller


Episode Details
Channel

All-In Podcast

Published

7/23/2025

Episode Summary

This installment of the 'Winning the AI Race' series on the All-In Podcast features in-depth interviews with four pivotal figures shaping America's technological and industrial landscape. First, James Litinsky, the CEO of MP Materials, explains his company's critical role in the domestic supply chain, transforming a bankrupt mine into the sole US producer of Rare earths. These materials are the essential feedstock for Physical AI technologies like robots and drones. Litinsky details the transformative MP Materials-DoD Deal, a landmark Public private partnership with the DoD to secure the US magnet supply chain, directly countering China's global dominance achieved through tactics like Chinese mercantalism. He notes that this deal, which also involves supplying major companies like Apple from facilities in Texas, was spurred by a mandate from the President Trump administration and serves as a potential blueprint for future industrial strategy. Next, Lisa Su, CEO of AMD, provides a nuanced perspective on the progress and challenges of Onshoring semiconductor manufacturing. She discusses the operational realities of their partnership with TSMC at the new Arizona facility, comparing it to established operations in Taiwan. Su emphasizes the explosive growth across the entire AI ecosystem, driven by insatiable demand for advanced compute like AMD's AI chips and GPUs. She reflects on AMD's competitive success against rivals like Nvidia and Intel and ponders the trajectory toward Superintelligence. The conversation then shifts to infrastructure with Chase Lochmiller, CEO of Crusoe. He introduces the concept of the Infrastructure of intelligence, arguing that data centers are evolving into AI factories. Crusoe embodies a Vertically integrated AI infrastructure model, building massive facilities in locations like Texas to meet the demand from Hyperscalers. Lochmiller highlights that the escalating Energy consumption for AI is the primary bottleneck to future growth, requiring massive new energy investments. Finally, Jensen Huang, CEO of Nvidia, offers a sweeping vision of the AI era. He portrays AI as the ultimate AI as a technology equalizer, empowering everyone with new creative and technical abilities. He discusses the relentless innovation at Nvidia, driven by technologies like the Hopper GPU and the foundational CUDA software platform, which are core to the American tech stack. Huang predicts a Multi-trillion dollar AI infrastructure buildout and praises the push for Reindustrializing the United States, a vision he associates with President Trump. He argues that the dominance of the American tech stack is paramount, even as powerful Open source models like DeepSeek emerge from China, as their global adoption reinforces US technological leadership.

Investment Ideas
Key Topics & People

Large cloud service providers that stand to gain an oligopoly through heavy AI regulation.

TSMC
TSMC
Organization

Taiwan Semiconductor Manufacturing Company, currently ranking just ahead of SpaceX in global market capitalization.

China
China
PoliticalEntity

Country producing competitive open source AI models.

CEO of Nvidia, referenced regarding the exponential future demand for electrical power due to AI computing.

Critical elements essential for high-tech manufacturing, where supply chains are heavily dominated by Chinese processing capabilities.

Taiwan
Taiwan
Location

An island nation heavily arming itself and shifting defense budgets, driving demand for materials.

Texas
Texas
Location

A US state highlighted for having a fragile, independent electricity grid prone to failure during cold weather.

The process of rebuilding the American industrial base, which requires massive infrastructure investments.

Apple
Apple
Organization

A massive technology corporation noted for creating trillions in market cap with relatively low capital intensity.

The concept that AI is standardizing and democratizing intelligence across enterprise outputs.

GPU
Technology

Graphics Processing Unit, the standard hardware for AI that Cerebras purposefully chose not to build.

Intel
Intel
Organization

Legacy chipmaker noted for losing market share during the cell phone compute transition.

AMD
AMD
Organization

Chipmaker mentioned as historically missing the mobile compute shift and competing in the AI silicon space.

Nvidia
Nvidia
Organization

Dominant AI chipmaker known for its GPUs, which Cerebras competes against.

Pratt's strategy of working with anonymous billionaires and private capital to rebuild Los Angeles structures.

DoD
PoliticalEntity

Department of Defense, involved in military procurement and AI deployment.

AI systems that understand and interact with the physical world, representing a $50 trillion industry.

CUDA
CUDA
Technology

Nvidia's proprietary software stack that serves as a massive strategic moat.

Freely available foundational AI models enabling localized, private computing for enterprises, challenging closed-cloud monopolies.

US President who delivered a State of the Union address emphasizing the Rate Payer Protection Pledge and implementing sweeping tariffs.

DeepSeek
Organization

An AI company mentioned as one of the frontier labs that will soon release a more powerful AI model trained on next-generation Blackwell servers.

Crusoe
Organization

A vertically integrated company that builds AI cloud infrastructure, including data centers, energy development, and application layers, with a focus on using stranded or sustainable energy sources.

CEO and co-founder of Crusoe, a company that builds and operates AI data centers powered by otherwise wasted or stranded energy sources.

The collection of technology platforms and services from US companies. Nadella believes its global success depends on broad adoption and creating ecosystem opportunities worldwide.

Arizona
Arizona
Location

The US state where TSMC is building its massive new semiconductor plant. It is cited as an example of the re-industrialization happening in America.

MP Materials
MP Materials
Organization

A key US-based rare earth company operating the Mountain Pass mine, which is part of the American effort to rebuild its domestic supply chain for these critical minerals with government support.

An aggressive economic strategy used by China, involving heavy government subsidies and market manipulation to dominate strategic industries like rare earths, effectively driving out international competition.

Superintelligence
Superintelligence
ScientificConcept

A hypothetical, advanced form of artificial intelligence. Johnson posits that its emergence will make human survival ('Don't Die') the most important virtue.

Lisa Su
Lisa Su
Person

The CEO of AMD, described as a 'total warrior' for making a 'bet the farm' decision on the deal with OpenAI to challenge Nvidia's market dominance.

Hopper
Technology

A high-performance GPU developed by Nvidia, used in large quantities to build AI supercomputers like xAI's Colossus.

A term used to reframe large-scale data centers for AI as facilities that are actively manufacturing intelligence, emphasizing their role as productive industrial assets rather than simple data processors.

The projection that building the global infrastructure for AI (factories, power, networks) will require trillions of dollars in capital investment over the coming years.

The specific strategic investment and offtake agreement between MP Materials and the Department of Defense, positioned as a new model for public-private partnerships.

A business model where a company controls multiple stages of the AI supply chain, from energy sourcing and data center construction to cloud services and operations.

The new foundational layer for the modern economy, comprising AI factories, energy generation, and high-speed networks necessary to power widespread AI.

The rapidly increasing demand for electricity to power AI data centers and training, which is becoming a major constraint on the growth of the AI industry.

The interconnected network of companies, technologies, talent, and capital required for the development and deployment of AI, from chip design to software applications.

AI chip
Technology

Specialized processors, like GPUs and custom ASICs, designed to accelerate artificial intelligence computations for training and inference.

The strategic initiative to relocate advanced semiconductor manufacturing facilities to the United States to enhance supply chain security and technological leadership.

Founder and CEO of MP Materials, a company he transformed from a bankrupt entity into the sole US supplier and refiner of rare earth materials.