The Circuit – Détails, épisodes et analyse

In this episode, Ben Bajarin and Jay Goldberg discuss NVIDIA's recent earnings and the challenges of managing expectations in the stock market. They touch on the mispronunciation of NVIDIA, the volatility of the stock, and the impact of whisper numbers on investor sentiment. They also discuss the role of NVIDIA's CEO, Jensen Huang, and the difficulty of managing investor relations. The conversation covers topics such as the slowdown in growth, gross margins, and the sustainability of high gross margins in the semiconductor industry. In this conversation, Ben Bajarin and Jay Goldberg discuss the reasons why they are not concerned about a bubble popping in the tech industry. They highlight the high demand for NVIDIA's products, the company's superior total cost of ownership (TCO) relative to competitors, and the continued growth of AI and data center infrastructure. They also touch on the potential for NVIDIA to expand its software business. The conversation then shifts to Intel, with the news that the company is exploring strategic options, including spinning off its Foundry division. Ben and Jay discuss the challenges of spinning off Foundry and the potential for a white knight investor to step in and support Intel.

Use code TC24 at wwic24.com

This week Ben and Jay are joined by AMD's EVP of Datacenter Solutions Forrest Norrod. AMD's acquisition of ZT Systems is a strategic move to enhance its position in the data center and AI market. ZT Systems is a leading systems design company focused on hyperscalers, with expertise in customized server and rack designs. By integrating ZT's design capabilities with AMD's silicon and software engineering, AMD aims to create better and more optimized AI server solutions. The acquisition also allows AMD to accelerate time-to-market and support their customers in deploying AI systems faster. AMD plans to divest the manufacturing side of ZT Systems and continue to collaborate with OEM and ODM partners.

Apple announced Apple intelligence, integrating AI in a very Apple way. They are focusing on on-device AI to align with their privacy philosophy. The features are only available on the most recent Apple Silicon iPhone pro and newer models. The implementation of AI is focused on useful feature-centric applications rather than general AI. Apple's infrastructure play with their own data centers and Apple Silicon has strategic implications and could potentially lead to them becoming a cloud service provider for developers. The closed-loop environment of security and privacy is a key advantage for Apple. The integration of client to cloud with a common software layer is a powerful trend. Other companies like Arm are also exploring this approach.

Computex has evolved into an important tech trade show, with a focus on gaming, GPUs, and components for gaming and PCs. The show has seen a shift towards data center components and AI servers, which have become a significant category. AMD and Intel are both making chips for AI PCs, with AMD leading in NPU tops. Qualcomm's Snapdragon is ahead in performance, spooking other vendors. The PC landscape has permanently changed, with more competition for Intel. The evaluation of product truth is crucial, as performance and battery life are key factors. In the data center space, AI servers are gaining attention and have higher content per dollar than general-purpose servers. However, market sizing for AI servers is challenging due to varying estimates and counting methods. The conversation explores the market potential for silicon sales to hyperscalers and tier two/tier three cloud service providers. It discusses the growth of the hyperscaler market and the potential for the tier two/tier three market to also grow. The conversation also touches on the challenges faced by new entrants in the market and the dominance of Nvidia in the AI server market. It explores the role of networking components in AI data centers and the margin opportunities for companies in the AI server market. The conversation concludes with a discussion on Arm's CSS service and its expansion into mobile and client markets.

With Special guests, Austin Lyons and Paul Karazuba! 

The conversation covers the topic of NPUs (Neural Processing Units) and delves into their architecture, performance, and relevance in the semiconductor industry. The discussion also addresses the use of tops per watt as a metric for NPU performance and the design origins of NPUs, including the use of licensed IP and the evolution from DSPs to NPUs. The conversation delves into the topic of NPUs (Neural Processing Units) and their integration into various devices. The speakers discuss the origins of NPUs, their design philosophy, and the potential impact on consumer devices. They also explore the role of NPUs in edge devices, AI PCs, and the future of consumer use cases for AI. The conversation concludes with predictions about the widespread adoption of NPUs and their potential impact on the market.

In this episode, Ben Bajarin and Jay Goldberg discuss the recent Google I/O event and the introduction of new TPUs and CPUs. They also explore the ARM architecture in the data center and the potential for switching between different ARM CPUs. Additionally, they touch on Qualcomm's launch of Windows on ARM and the significant adoption of Qualcomm SKUs by major OEMs. The conversation covers various topics including the dynamics in the PC space, Nvidia's earnings, and the 10-to-1 stock split. The main takeaways include the strong position of Windows in the PC ecosystem, the growth of Nvidia's data center networking business, and the potential impact of a dedicated inference chip. The conversation also discusses the competition in different parts of Nvidia's stack and the volatility that may come with the stock split.

In this episode of The Circuit, Ben Bajarin interviews Esam Elashmawi, Chief Strategy and Marketing Officer of Lattice, about the world of Field-Programmable Gate Arrays (FPGAs). They discuss the basics of FPGAs, their unique capabilities, and their pervasiveness across various applications. They also explore the advantages of FPGAs over Application-Specific Integrated Circuits (ASICs) and the flexibility they offer in terms of customization and reprogramming. Esam highlights the role of FPGAs in different markets, such as communications, computing, industrial, and automotive, and how Lattice differentiates itself in the FPGA market. They also touch on the challenges of building an FPGA company and the potential of FPGAs in AI applications, both in data centers and at the edge.

In this episode of The Circuit, Ben Bajarin and Jay Goldberg interview Aart de Geus, the executive chair and co-founder of Synopsys, about the semiconductor industry and the evolution of electronic design automation (EDA). They discuss the concept of EDA and how it allows designers to simulate and optimize circuits without physically building them. Aart de Geus highlights the shift from analog to digital design as a major inflection point in the industry. He also explains the role of EDA tools in automating the design process and the importance of functional correctness in chip design. The conversation then delves into the future of chip design, including the use of AI in EDA tools, the rise of chiplets and vertical stacking, and the increasing complexity of system-level design.

In this episode of The Circuit, Ben Bajarin and Jay Goldberg discuss the latest developments in the semiconductor industry, focusing on the sectors of data centers, PCs, mobile, automotive, and industrial semiconductors. They analyze the performance of companies like Intel, AMD, Nvidia, Apple, Qualcomm, and NXP, and explore the factors influencing their success or challenges. The conversation highlights the impact of AI on the industry, the dynamics of the Chinese market, the potential for a super cycle in PCs, and the outlook for the automotive and industrial semiconductor sectors.

Summary

In this episode, Ben Bajarin and Jay Goldberg discuss the skyrocketing capex by cloud hyper scalers and its implications. They explore the significant growth in capex spending by companies like Google, Microsoft, and Amazon, and the connection between capex and revenue growth. They also discuss the challenges of monetizing AI investments and the need for clear consumer use cases. The conversation touches on the potential for new innovations in AI wearables, AR glasses, autonomous cars, and robotics. They also consider the deflationary effects on costs and the timeline for the development of AI applications.