The Circuit

Unpacking Marvell's Analyst Day, OpenAI Revenue Rumors, and the Heterogeneous Rack

Has the Memory Cycle Broken? Plus: Nvidia & Broadcom’s $40B+ Financing Playbook

The semiconductor and AI infrastructure sectors are undergoing a fundamental transformation characterized by shifting market cycles and new financial strategies. Memory manufacturers face a debate over whether AI-driven demand and sustained bit growth have permanently muted historical cyclicality or if capacity expansi...

Memory Pricing Pressures, Meta's Muse AI, and the Agentic CPU Crunch

The memory market faces significant shifts as Chinese manufacturers like CXMT and YMTC accelerate capacity expansion, pressuring global pricing and forcing OEMs to navigate complex storage and memory tiering strategies. Meanwhile, the emergence of AI agents—such as Meta’s new assistant—highlights a critical transition ...

Conversation with Ayar Labs CEO Mark Wade on Silicon Photonics, Co-Packaged Optics, and Next-Gen AI Infrastructure

Large-scale AI data centers face a critical revenue-scaling bottleneck caused by the performance limits of electrical interconnects in compute systems. Optical connectivity offers the only viable solution to overcome the bandwidth-distance trade-off, enabling the next generation of token throughput and reasoning model ...

EP 192: Talking Apple Silicon with Apple's Tom Boger, Kaiann Drance, and Sri Santhanam

Apple’s vertical integration strategy drives product innovation by aligning silicon architecture with hardware and software development. The transition to a 2-nanometer process node enables significant performance gains, including doubled neural engine capacity and increased memory bandwidth, which are essential for ha...

Ep 191: Agents Gone Wild: Broadcom’s ASICs and the AI Frontier

The AI infrastructure market is undergoing a complex transformation as custom ASIC demand surges and supply chain bottlenecks persist. Broadcom’s recent earnings highlight a shift where hyperscalers like Anthropic and OpenAI are becoming dominant customers, though modeling this market is increasingly difficult due to d...

EP 190: NVIDIA and Marvell Earnings, Hot Chips Hot Takes

NVIDIA’s recent earnings report signals continued growth, with 70% revenue guidance for the coming year despite ongoing supply chain constraints. While NVIDIA remains the dominant player, the industry is entering an "era of ferment" regarding inference architecture, characterized by diverse technical approaches to sili...

EP 189: Analyzing ADI Earnings, Custom Silicon Shifts, and Semiconductor Debt

Analog semiconductor companies are increasingly integrating into the AI data center ecosystem, specifically through high-spec power management systems for 800-volt architectures. Analog Devices reports that data center revenue is doubling as it addresses new power conversion and safety requirements. While AI sentiment ...

EP 188: The Wild West of Optical, Cisco's AI Hack, and Is Compute an Investable Asset

The AI infrastructure build-out faces significant technical and economic challenges, particularly regarding the transition to advanced optical networking and the sustainability of massive capital investments. While near-packaged optics (NPO) gain traction, co-packaged optics (CPO) remain a distant, manufacturing-intens...

EP 187: Inside the Future of Memory Summit & AMD’s "Boring" Bullishness

AMD’s recent earnings underscore a disciplined execution strategy, prioritizing long-term roadmap commitments over short-term market volatility, particularly within the data center CPU segment. The industry is reaching a critical inflection point for Compute Express Link (CXL) technology, which facilitates memory disag...

EP 186: Mobile Bottlenecks, Skyrocketing AI CapEx, and the Compute Crunch

The mobile market faces persistent weakness, with demand shifting toward lower-tier devices and supply chain constraints hindering production for major players. Meanwhile, the AI-driven data center sector experiences massive capital expenditure growth as hyperscalers prioritize inference-bound workloads. Amazon emerges...

EP 185: AMD Advancing AI, Intel Earnings and WFE Demand

The semiconductor industry is undergoing a structural shift toward rack-scale compute, characterized by intense collaboration and competition among major players. AMD’s recent AI event highlights this evolution, with a focus on inference-optimized hardware and strategic partnerships like Anthropic to improve token effi...

Ep 184: Earnings TSMC/ASML/AEHR, State of AI Semis

Semiconductor industry growth is currently defined by an aggressive capital expenditure cycle driven by the transition from monolithic smartphone chips to complex AI and high-performance compute architectures. TSMC and ASML are scaling production and increasing prices to meet this demand, reflecting a fundamental shift...

EP 183: SKHynix IPO, AI Model Wars, and NVIDIA Supply Chain Rumors

SK Hynix’s recent US IPO highlights the growing necessity for global semiconductor firms to access deeper capital pools and diversify shareholder bases amid a sustained memory supercycle. While memory remains inherently cyclical, the shift toward hyperscaler-driven demand suggests long-term scarcity, fueling ongoing de...

EP 182: AI Fatigue , Meta's Neo-Cloud Rumors , and Samsung's Foundry Comeback

Semiconductor market volatility stems from a disconnect between speculative investor sentiment and underlying fundamentals. While AI-driven capital expenditure remains high, companies like Meta face pressure to justify massive infrastructure investments, leading to strategic shifts such as potential capacity leasing to...

EP 181: Cerebras Earnings, QCOM Investor Day, Micron Earnings and the Memory Mafia

The tech industry faces a shifting landscape as companies navigate the high-stakes demands of AI infrastructure and memory supply. Cerebras’ initial public earnings highlight the complexities of scaling wafer-scale engines, where high demand for fast inference clashes with margin pressures and unique architectural cost...

EP 180: HPE Reflections on AI Compute Racks, Analog Analog Analog Semis.

AI infrastructure is undergoing a radical transformation as OEMs like HPE and Dell move beyond standard reference designs to address the extreme power and thermal demands of modern AI workloads. By engineering specialized hybrid-liquid cooling systems and high-density racks, these companies are optimizing compute-per-m...

EP 179: Apple's WWDC and Siri AI, Nebius Inflection, China Trip Takeaways

Apple’s recent Worldwide Developer Conference signals a shift toward "applied AI," prioritizing system-level integration and semantic search over standalone chatbot features. This strategy relies on rebuilt indexing and a commitment to privacy, while marking a notable pivot in Apple’s infrastructure strategy through th...

Ep 178: COMPUTEX! Broadcom Earnings, other nuggets from the Week

Computex has shifted from a showcase for consumer PC hardware to the epicenter of AI-driven data center infrastructure, characterized by massive server deployments, liquid cooling, and 800 VDC power requirements. NVIDIA’s entry into the notebook CPU market via the RTX Spark collaboration with MediaTek signals a broader...

EP 176: Earnings! Marvell, Synopsis, Dell and more thesis time.