The Circuit

They unpack Nvidia's GTC buzz, product launches, and how vertical stacks reshape data center infrastructure. They explore optical networking trends, co-packaged optics reliability, and manufacturing challenges. They react to Micron's surprising earnings and debate whether memory margins signal a lasting industry shift.

In this episode of The Circuit, hosts Ben Bajarin and Jay Goldberg dive into NVIDIA’s strategic positioning within the AI infrastructure stack, colloquially referred to as a "five-layer cake". They dissect NVIDIA’s recent push into open-source models, comparing it to Intel’s historical dominance through Linux optimizations, while questioning the true intent behind NVIDIA’s $26 billion "investment" in open-source R&D. The conversation shifts to the rapidly evolving world of "agentic orchestration," where AI agents collaborate across different models to maximize efficiency and token spend. Finally, the duo explores the broader hardware ecosystem, including Meta’s ambitious six-month silicon cadence and the supply chain "masters" who hold leverage in a world defined by compute constraints.

Analysis of Broadcom and Marvell earnings and what their guidance signals for AI semiconductor demand. A deep dive into the copper versus fiber debate for data center interconnects and when optics become necessary. Discussion of hyperscalers' varied networking strategies and the supply chain pressures shaping compute availability. Recap of Apple's new affordable Neo and M5 product moves and their market implications.

A whirlwind week of semiconductor and AI news, from AMD's deal with Meta to NVIDIA's big quarter and market reaction. They unpack supply constraints, multi‑vendor strategies, and memory and networking plays. Funding moves by OpenAI and AWS's role get attention, along with neo‑cloud economics and rising revenue at Anthropic. The conversation touches on AI agents, SaaS disruption, and why humans may still matter.

A discussion on tariffs, supply chain strains and who absorbs rising costs. They map simultaneous shortages across GPUs, memory, substrates and equipment. The conversation covers power and cooling limits shaping AI data centers and the move from copper to fiber and photonics. They flag hyperscaler orders as early demand signals and consider potential two-track global supply chains and M&A shifts.

They dig into Nebius and CoreWeave financial moves and what rising contracted wattage means for neocloud growth. They debate hyperscalers becoming heavy-asset utilities and the forecasting risks that follow. They cover Applied Materials’ wafer equipment surge and memory tightness reshaping supply and pricing. Networking vendors’ fate amid hyperscaler verticalization also gets a sharp focus.

A deep dive into the hyperscaler CapEx surge and how massive budgets are reshaping cloud competition. Conversation about who wins from big compute buys, from GPUs to custom silicon. Discussion of semiconductor pressures from rising memory and weaker handsets. Analysis of strategic M&A moves in timing and analog markets and what that means for the industry.

Discussion of a semiconductor WFE boom and why production constraints cap revenue upside. A look at hyperscalers' differing AI CapEx strategies and internal compute battles. Debate over whether massive AI compute spending will show returns. Examination of storage and memory market volatility and the tug-of-war in supplier contracts. Consideration of analog industry inventory shifts and long-term supply scarcity.

They dig into Intel's strong Q4 but cautious guidance tied to wafer capacity constraints. Listeners hear debate over foundry credibility, advanced packaging yield challenges, and whether Intel can win third‑party customers. The conversation also covers CPU demand for AI, ARM gaining server share, and how hyperscaler design choices shape the datacenter landscape.

The hosts dive into TSMC's recent earnings, unpacking the complexities of semiconductor cyclicality and the challenges posed by advanced packaging. They explore TSMC's cautious CapEx strategies amid rising AI demand and discuss the intricate relationship between wafer demand and manufacturing capacity. The conversation then shifts to OpenAI's partnership with Cerebras, examining new monetization efforts like ads and subscription tiers to address compute costs. They also analyze the fluctuating GPU pricing landscape and the implications for hyperscalers and enterprise clients.