Chipstrat

Co-Packaged Optics (CPO) Primer - The transceiver disappears. Optics move onto the switch package - Who captures value as the optical supply chain reshuffles - EMLs and DSPs give way to silicon photonics and external CW lasers - Meta seeing fewer failures with CPO than pluggables Link: chipstrat.com/p/optics-prime…

SemiDoped: @MicronTech, MicroLEDs, and AI - Jensen's $250K Token Budget Per Engineer - On-Prem Inference vs. Cloud Token Spending (Dell Pro Max, CapEx vs OpEx) - @MicronTech Earnings & HBM4 in Vera Rubin - Why AI Drives a Step Change in Memory Demand - Micro LEDs – The "Slow but Wide" Approach - VCSELs & @Lumentum's 1060nm Scale-Up Play $LITE $MU youtube.com/watch?v=qD4Lk2…

Three things from Jensen’s GTC 2026 keynote were notable: - Scale-up is going copper and optical - Nvidia is selling lots of standalone CPUs, calling it a multi-billion-dollar business - The Groq 3 LPX chip is real, fabbed by Samsung and shipping Q3 We cover the full Vera Rubin data center system, tiered inference economics, CPO in production, and take a detour into whether the vibe-coding era actually kills SaaS (we think not). chipstrat.com/p/gtc-2026-key…

Meta's MTIA Roadmap: - Further validates the industry is past “GPUs for everything” value prop - Implications for Broadcom, Nvidia, AMD, HBM suppliers, TSMC, Arista, and inference startups chipstrat.com/p/metas-mtia-r…

Whiplash week. Optics? Copper? Both? Mon AM: Nvidia bets $4B on optics. Mon PM: Credo posts 200% YoY growth on copper. Wed PM: Hock Tan claims 400G/lane works over copper, potentially pushing CPO past 2030. 48 hours of whiplash. Optics? Copper? The answer is both. The question is when. This week @theaustinlyons and @vikramskr unpack: - Nvidia locking up laser supply - Credo’s blowout quarter and the reliability thesis - Broadcom’s copper bombshell - A 4D chess theory on why Hock Tan downplays optics when Broadcom is a CPO company Chapters (00:00) - Newsletter Plugs: Groq LPUs & Broadcom’s Laser Business (03:15) - Dynamo & the Rise of Workload-Specific Hardware (08:04) - Austin’s Broadcom Laser Deep Dive (09:53) - The Week’s Whiplash: Optics Monday, Copper Wednesday (17:50) - Why Nvidia Invested $4B: Geopolitics, Supply & the HBM Playbook (24:15) - CPO Lasers & Optical Circuit Switches (26:16) - Credo Earnings: 200% YoY Growth & the Copper Bull Case (31:09) - Reliability, AECs & Oracle’s GPU Cluster Problem (35:48) - Credo’s Optics Play: Micro-LED Active Cables & the CPO Timing Risk (38:45) - Broadcom Earnings: Hock Tan’s Copper Bombshell (43:34) - Customer-Owned Tooling: Hock Tan Says “Good Luck” (44:25) - Vik’s 4D Chess Theory: Why Hock Tan Talks Up Copper (47:03) - Wrap-Up: It’s Both — The Real Question Is Timing $AVGO $CRDO $NVDA

Broadcom Makes Lasers? ($) chipstrat.com/p/broadcom-mak… $AVGO $LITE $COHR

Lumentum and the Laser Bottleneck ($) chipstrat.com/p/lumentum-and…

Block built 3nm Bitcoin mining rigs and made some sweet fleet management software too. Does it matter? chipstrat.com/p/jack-dorseys…

"Block is building a 3nm Bitcoin mining ASIC. Huh? A financial company designing hardware on leading-edge process technology? This news led me down a rabbit hole of research, and I’m sharing what I learned in a two-part series. To grasp Block's Bitcoin strategy, we first need to understand Block’s core businesses and hardware strategy. Part one will provide this essential context, while part two will explore their specific Bitcoin endeavors." chipstrat.com/p/hardware-as-…

What if fabs bought photons the way they buy process gases? chipstrat.com/p/photons-as-a…

$ANET bet on Ethernet. Good bet. Ethernet TAM is expanding, and agentic AI helps. Priced in though? chipstrat.com/p/aristas-3-bi…

The optical networking supercycle is here! In this podcast, @theaustinlyons and @vikramskr go through all the tech jargon and explain what everything means. In just about 45 mins, you will know everything required to keep up with the next revolution in AI. Chapters 00:00 Introduction to AI and CPU Bottlenecks 03:00 The Rise of Silicon Photonics 06:01 Understanding Optical Networking and Data Centers 08:49 Scale Across: Connecting Data Centers 11:56 Scale Out: Optimizing Data Center Connectivity 14:53 Scale Up: The Future of GPU Connectivity 23:32 The Shift from Copper to Optical Connections 26:13 Challenges and Reliability of Lasers 30:47 Understanding Co-Packaged Optics 34:17 Market Dynamics: Demand and Supply of Lasers 40:46 Emerging Technologies: Optical Circuit Switches

Hyundai alone signed a deal for 50K Blackwell GPUs for its autonomous driving data center. At ~$30-40K/GPU, that's $1.5-2.5B from a single automotive OEM. That's roughly equal to Nvidia's entire reported annual automotive revenue. But it doesn't show up in automotive. It shows up in the $NVDA Data Center business. chipstrat.com/p/the-nvidia-b…

Meta doesn't need the smartest AI. It needs the most profitable one. GEM has unlocked LLM-like scaling laws for ad ranking. The strategy is simple: GEM now, LLMs later. New post: chipstrat.com/p/metas-roic-s…

Yes for @chipstrat!

My analysis of $AMD's $10B quarter and stock selloff: - MI450 timing & OpenAI - inference portfolio vs. Nvidia - server CPU - and the "why now?" investor problem chipstrat.com/p/the-mi450-wa…

New Episode: OpenClaw Makes AI Agents and CPUs Get Real Vik explains ways he uses Claude Code not to Code. But instead, to do research and keep up to date with semis. Plus ... → AI agents are great, but does everybody need them? → Deploying OpenClaw safely and not losing like a crypto-bro → AI agents cannot understand "what is not said" → Becoming more self-aware for AI agents to work Chapters (00:00) Introduction: Why 2026 may be the year of AI agents (01:12) What people mean by agents and the OpenClaw naming chaos (02:41) Agents behaving badly: crypto losses and social posting (03:38) Claude Code as a research tool, not a coding tool (05:54) Terminal-first workflows vs GUI-based agents (07:44) Connecting Claude Code to Gmail, Drive, and Calendar via MCP (09:12) Token waste, authentication friction, and workflow optimization (10:54) Automating newsletter ingestion and research archives (12:33) Giving agents login credentials and security tradeoffs (13:50) Filtering signal from noise with topic constraints (16:36) AI-driven idea generation and its limitations (17:34) When automation effort is not worth it (19:02) Are agents ready for non-technical users? (20:55) Why OpenClaw should not run on your personal laptop (21:33) Safe agent deployment: VPS vs local servers (23:33) The true cost of agents: infrastructure plus inference (24:18) What OpenClaw adds beyond Claude Code (26:53) Agents require managerial thinking and self-awareness (28:18) Local inference vs cloud APIs (30:46) Cost control with OpenRouter and model hierarchies (32:31) Scaling agents forces model and cost optimization (33:00) AI aggregation vs creator analytics (35:58) AI as discovery, not a replacement for reading (38:17) When summaries are enough and when they are not (39:47) Why AI cannot understand what is not said (41:18) Agentic AI is driving unexpected CPU demand (41:49) Intel caught off guard by CPU shortages (44:53) Security, identity, and encryption shift work to CPUs (46:10) Closing thoughts: agents are real, early, and uneven @vikramskr @theaustinlyons

Microsoft 🤝 Ethernet scale-up networking. The philosophy - open, multi-vendor Ethernet - standardized switches - differentiation pushed to the transport layer

Intel Q425: Back to Reality ($) - Continued supply constraints - Advanced packaging timing - Why Q1 2026 will be more of the same chipstrat.com/p/intel-q425-b…