In the past 30 days, @alexandr_wang has posted only twice on Meta’s Threads, versus 25+ times on @X.
Hey @nikitabier and @elonmusk, can you enable creator monetization for Alexandr?
- Just as Zuck’s Instagram copied Snapchat Stories and other major Snapchat features, Mark Zuckerberg is now copying @elonmusk’s GPU cloud strategy: launching a large-scale cluster on short notice with a huge pricing premium. We break it down below👇️ 1/3🧵Elon Musk shocked the AI infrastructure world when he announced the first Anthropic deal. He shocked it again when he landed the Google deal. The revenue per MW from Elon’s deals is roughly triple and quadruple what peers are charging, respectively. The profit-per-MW gap is, ofOur Datacenter Model breaks down Meta’s and xAI’s capacity additions quarter by quarter across self-builds, datacenter leasing, and cloud renting. Our newsletter article breaks it down too. 3/3🧵
- ALERT: On vLLM Kimi, which has the same model architecture as xAI’s Cursor Composer 2.5, NVIDIA mogs AMD. Also great to see B300 faster than B200.
- We imported more from Taiwan than China as tariffs and AI trends collide. (1/6)🧵Replying to @SemiAnalysis_But chip imports from Taiwan are fundamentally different from tube socks from China. Chips have an enormous amount of US IP value embedded in them that gets added back in when GDP is calculated. (5/6)And crucially, chip imports are building the US capital stock — presumably increasing future GDP as AI capabilities expand. (6/6)
- QUESTION MARK: One of Elon’s companies has limited employee token usage to just $200 per week, while another Elon company is saying everyone will use so much AI that we’ll need space data centers because Earth will run out of power, plus TeraFabs to produce a terawatt of chips
- Correct call by NVIDIA tbh. Why “waste” wafers and GDDR on fun when Rubin pays triple? So gamers get the only GeForce cards the margin math allows: the paper ones. Enjoy the mealworm buffet, fun is a datacenter-exclusive feature this gen, and the rest of us are getting GeForced
- The coherent-lite transceiver technology was presented by Google in 2021 for an IEEE meeting as an alternative to the shorter-reach, traditional CWDM4 IM-DD transceiver, and we are finally expecting to see volume adoption at Google. As flagged in May to our institutional clients,The bottom line is this: even though we expect co-packaged optics (CPO) to eventually be the go-to bandwidth scaling solution for optics, scale-up CPO will have to work through the integration, yield, and ecosystem maturity curve before it is ready for volume manufacturing and
- The Future of Meta Superintelligence: A 1 Year Progress Update A top tier RL environment startup spawns out of thin air, the most aggressive compute ramp we've ever seen, 2000km+ scale-across, and some advice for Google DeepMind
- Vera Rubin NVL72 is a beast, thanks to its extreme co-designed system architecture, and First At-Scale Production Token (FASPT) is fast approaching. Below, we break down the architecture and some of the latest manufacturing innovations.👇️ 1/7🧵Replying to @SemiAnalysis_The compute tray topology of VR NVL72 is roughly similar to that of GB200 and GB300, though there are some differences. 6/7🧵We do a full breakdown of the BoM and power budget in the VR NVL72 Component BoM Model. It provides a system-level bill of materials and power budget analysis for the VR NVL72 system. 7/7🧵
