The Circuit

In this action-packed episode of The Circuit, Ben Bajarin and Jay Goldberg recap a hectic week in Silicon Valley, spanning Semicon West in San Francisco, the Open Compute Project (OCP) Summit in San Jose, and major analyst events. The hosts examine the evolving semicap and wafer equipment landscape—including insights from ASML, TSMC, and Applied Materials' new EPIC Center launch—noting how toolmakers are re-engineering cleanroom capabilities and pushing reticle limits to support massive AI accelerators and advanced packaging. They then transition to the OCP Summit floor, unpacking the physical and engineering realities of modern data center infrastructure. The discussion covers the immense crowds surrounding liquid cooling innovations (such as Delta Electronics' 3MW manifold system), the shift toward 800V DC power architectures, and the complex choices data center operators face when balancing co-packaged optics (CPO) against the surprising longevity of copper interconnects. Finally, Ben and Jay analyze the architectural trade-offs of scaling compute domains across multi-rack topologies versus squeezing maximum GPU density into single racks.

In this episode of The Circuit, Ben Bajarin and Jay Goldberg kick off a packed conference week (Semicon West, Applied Materials' EPIC Center launch, and the sold-out OCP Summit) before diving into Marvell's upbeat Analyst Day in New York. They examine how Marvell has shifted its narrative beyond Amazon custom ASIC share battles toward a broader data center infrastructure play—spanning optical interconnects, switching silicon, and a newly teased class of AI infrastructure management silicon built to work alongside traditional BMCs and FPGAs in complex, heterogeneous scale-up compute domains. Next, the hosts unpack the recent market panic over leaked OpenAI ARR figures, arguing that the episode exposes the fragility of the AI semiconductor trade and how physical infrastructure bottlenecks—specifically whether the U.S. grid can energize 20 to 30 gigawatts of power next year—pose a much greater risk to chip timelines than end-user demand. Finally, they close with a strategic theory on why TSMC is granting AMD extra wafer capacity to potentially throttle Intel's cash flow as Intel Foundry becomes a credible threat, and how a looming 20% downturn in the PC market could ironically free up critical wafer supply for higher-margin server CPUs and custom AI ASICs.

In this episode of The Circuit, Ben Bajarin and Jay Goldberg unpack the shifting structural economics of the semiconductor and AI landscape. Following Micron’s latest earnings, they kick off with a deep-dive debate on whether the memory market has permanently raised its floor to $600B–$800B, or if rapid capacity expansion from players like China’s CXMT will trigger a classic cyclical reckoning by 2028. They also examine the trade-offs between high-margin HBM and "trailing-edge" legacy memory like DDR4, as well as how rising component costs are impacting consumer PC and smartphone upgrade cycles. Next, Ben shares on-the-ground insights from HPE’s investor event, questioning whether HPE's post-Juniper pivot truly makes it a networking company as enterprises scramble to upgrade aging infrastructure against agentic AI security threats. The hosts then analyze the rise of "Balance Sheet as a Service"—breaking down Nvidia’s bold vendor financing disclosures and Broadcom’s reported $42 billion loan commitment to Anthropic ahead of its IPO, and why fabless chip designers are taking on this massive capital risk instead of foundries like TSMC. Finally, they explore "Tokenomics," looking at how investors and enterprise buyers can move past AI hype to quantify the hard-dollar unit economics of input/output tokens, hardware optimization, and specialized frontier models

In this episode of The Circuit, Ben Bajarin and Jay Goldberg explore the evolving dynamics of the semiconductor memory market, consumer AI agents, and critical compute infrastructure. The hosts begin by analyzing the aggressive capacity expansions and yield improvements of Chinese memory providers YMTC and CXMT, discussing how their growing output might provide pricing relief for consumer electronics OEMs currently squeezed by massive data center demand. The conversation then pivots to Meta's new "Muse" AI assistant, which the hosts view as a potential catalyst for mainstream AI adoption that could bridge the user skill gap by acting as a proactive personal and enterprise orchestrator. Finally, they unpack Anthropic's $11 billion equity-exchange partnership with Akamai, highlighting its implications for distributed edge inference and the massive, unforeseen surge in CPU demand required to power these continuous "agentic" workloads.

In this episode of The Circuit, hosts Ben Bajarin n and Jay Goldberg speak with Mark Wade, CEO of IR Labs, and Jesse Leiter, VP of Capital Strategy and Investor Relations. The conversation focuses on the necessity of solving the interconnect bottleneck in large-scale AI data centers by transitioning to optical connectivity. Mark explains the immense complexity of integrating optical engines into advanced CMOS nodes and highlights how the release of ChatGPT accelerated the market's demand for these large-scale computing solutions. Additionally, the discussion explores IR Labs' strategic reliance on leading-edge foundries like TSMC, their ecosystem partnerships with tech giants like NVIDIA, AMD, and Intel, and their calculated funding strategy aimed at an eventual IPO in a market projected to reach $30 to $50 billion by 2030